9 replies to this topic

[nasim303]

hello,

 

I want to know how to add overlapping cuts in Petroleum Assays in ASPEN HYSYS V11. 

 

Thanks in advance.

Attached Files

•  hysys input option.PNG   476.43KB   5 downloads
•  assay sample.PNG   194.71KB   4 downloads

Edited by nasim303, 18 January 2023 - 03:58 AM.

[Pilesar]

If I understand your question accurately, you just add in the measured products as separate streams to the model and then flow them together in the flowsheet to combine them.

[nasim303]

If I understand your question accurately, you just add in the measured products as separate streams to the model and then flow them together in the flowsheet to combine them.

Thanks for your kind concern. But actually I meant to say that when inputting the assay data (like, ASTM D86 or other bulk properties), a whole crude and multiple cut options there are to input in HYSYS>>Properties>>Petroleum Assays (screenshot attached in my main post) . But the problem is for each consecutive cut the IBP is equal to the FBP of the immediate previous cut! But practically there are overlapping in the cuts of the crude. So How we can input those data? (I shared a sample of overlapping crude in my main post.)

Edited by nasim303, 18 January 2023 - 04:00 AM.

[nasim303]

My another question is when I characterize the assay after inputting the data why some manually inputted data is manipulated by HYSYS?

is it possible to keep all manually inputted data unchanged?

[Pilesar]

You should have many more than 5 petroleum cuts in your assay for pseudocomponents. A good range to start would be boiling points no greater than every 15 C apart. Assay component distribution is not the same as crude product ranges. There will be overlapping boiling points in the crude column products. Without enough petroleum cuts in your assay range, you will not get useful results in your model. The material that boils higher than your heaviest defined assay cut gets combined into the last cut and this will change your input. Your distillation curve input shows 'loss' and 'residue' which acknowledges that the IBP and FBP do not reflect the boiling points of the lightest and heaviest components. The simulation software is fitting your stream input to a curve in order to characterize it according to the assay pseudocomponents you requested.

[SilverShaded]

Enter the data as a single stream property.  The hypo-compents need to be allready in your componenet list, an easy to do that is add an assay and it will ask you which list of hyps to use.  Don't use the list that goes to 1150C because it's total nonsense, use the 850C one.

 

IBP and FBP are also nonsense usually, they should be entered as 1% or 99% or estimated using a probabity plot, or just leave them out and Hysys will estimate using a probability plot.

Edited by SilverShaded, 18 January 2023 - 09:34 AM.

[nasim303]

You should have many more than 5 petroleum cuts in your assay for pseudocomponents. A good range to start would be boiling points no greater than every 15 C apart. Assay component distribution is not the same as crude product ranges. There will be overlapping boiling points in the crude column products. Without enough petroleum cuts in your assay range, you will not get useful results in your model. The material that boils higher than your heaviest defined assay cut gets combined into the last cut and this will change your input. Your distillation curve input shows 'loss' and 'residue' which acknowledges that the IBP and FBP do not reflect the boiling points of the lightest and heaviest components. The simulation software is fitting your stream input to a curve in order to characterize it according to the assay pseudocomponents you requested.

I got your answer for the last part. The red color part answer is not clear to me unfortunately. Can you please make me this understand again?

[Pilesar]

A petroleum crude stream is made up of hundreds (maybe thousands) of chemical components.  Each component has its own characteristics such as molecular weight, boiling point, density, viscosity, etc. The composition is constantly changing from second to second as it emerges from the ground. A lab test to accurately identify each component and its quantity would be very expensive and soon out of date. Even if you had a completely accurate composition of these hundreds of components, not all of them have been isolated to determine their individual properties. To use this information, a simulation model would need the component properties in its database and the user would have to input the composition with hundreds of components. Each component in a simulation adds to the computational load.

 

So in practice, crude streams are not modeled by individual component except possibly at the lowest boiling end of the stream (the light ends.) Instead, useful information about the crude stream is found by boiling off a little bit of the stream at a time and measuring the properties of the portion of stream (or cut) that is boiled off.  The computer model takes each of these cuts and treats it as if it were a single component (a pseudo component) instead of the many components actually in the cut.  Because petroleum from the ground forms a natural curve of boiling points, the properties of the individual cuts (pseudocomponents) can be estimated from a few measurements of the entire stream.

 

You have seen that the products of the distillation have overlapping components. To model a distillation, your model will need to have several pseudocomponents in those overlapping sections. Because your assay is characterized with only five cuts total, there will be no way to accurately model the split between product streams.

 

If you get a converged solution, go back and increase the number of cuts in your assay and you will get better accuracy. The more assay cuts, the better you will be able to model the product split. Once you have enough cuts (with several pseudocomponents in the overlap sections) you will not gain a lot by adding more cuts. A cut every 10C to 15C should be good enough.

 

You will never match your experimental data exactly with a simulation model. The quandary is 'how close is good enough?' There are errors in measurements, errors in thermodynamics, errors in choice of assay characteristics, etc. Setting the number of cutpoints for the assay is fundamental to even have a chance to get close.

[SilverShaded]

My another question is when I characterize the assay after inputting the data why some manually inputted data is manipulated by HYSYS?

is it possible to keep all manually inputted data unchanged?

Hysys smooths out the curve using a probability plot.  So some distillation numbers may get changed.  Decades ago this was often done manually using probability graph paper.

[SilverShaded]

Just to add to some previous replies.  Occasionally Naphtha regions are modelled with real components using GC analysis data.  Simulators like Hysys do have all the components in the database to do this and the number of components in the Naphtha region varies by crude but can be upto around 200.  Unfortunatley GC data does identify mercaptans or thiophenes so it takes a bit of creatviity to include the sulphur curve in the Naptha region.