Black oil model vs. Compositional model?

[ginsoakedboy]

Can anyone please explain what do the following terms mean?

1. Black oil model (or modeling)
2. Compositional model

What are the differences between the two? Is one better than the other in any respect?


FYI:
I am trying to formulate an approach to analyze a portion of subsea production system in OLGA. I don't have any particular composition of produced fluid to consider, rather I'd like to generate a "representative" model.

 

[ginsoakedboy]

bump ...

 

[BigInch]

Always model only enough to get the answers you need.

I'd use black oil if you don't have to track composition for any reason, such as when batching one field's oil production behind another field's, each going to different delivery points or if flows must be blended to achive a manageable viscosity of two different oils etc. Compositional is usually a whole lot more input (error) intensive, slower to run and you get a whole lot more answers than you need to actually get the ones you're looking for.

Besides, you'll be lucky to get the compositional data you need anyway.

http://virtualpipeline.spaces.msn.com

"What gets us into trouble is not what we don't know, its what we know for sure" - Mark Twain

 

[ColourfulFigsnDiags]

Yes and No BigInch, a compostional model (if you have the composition data available) will generally provide more accurate results for PVT data than a black oil model based on a defined GOR and WC.

If you have compositional data available, I would use it. PVTSim or another simulator will produce a much better and more accurate phase envelope and fluid properties with it.

But judging by your FYI, use the black oil model, as trying to develop your own generic composition can be tedious.

Read the Eng-Tips Site Policies at FAQ731-376

 

[BigInch]

Well of course, but the nature of the question suggests that the accuracy is not needed at this time. If there was one thing that irritated me to no end about hydraulic consultants is that there was always somebody ready to cost me time and money to build a model "so accurate" and so detailed with each 10 ft length of wall thickness change and all the bends and fittings in each pump stations on 1500 miles of pipeline to arrive at answers that wern't any more accurate than 10% anyway.

I only suggest to save your time and your client's time for things that are really needed.

http://virtualpipeline.spaces.msn.com

"What gets us into trouble is not what we don't know, its what we know for sure" - Mark Twain

 

[ginsoakedboy]

Thanks for your answers, BigInch and ColourfulFigsnDiags. I also think it would be better to start with the black oil model. I am planning to use field specific compositions once the OLGA model is shown to be working.

Just for my curiosity, is there anything like a generic composition for GoM? Am I even justified in looking for a generic fluid be it any region?

 

[BigInch]

I think the answer really depends on what you are simulating, specialized production process, treatment or transportation. I would certainly agree that in some cases you would need a compositional model as I have explained above, but I would also think those cases where you do really need a compositional model will more rare than common. In other words, really not necessary to have a template already available.

There is a lot of information about various crudes from around the world, but the info is more physical than chemical in nature and lists such things as SG-API gravity, vapor pressure, viscosity and maybe a heat capacity. The rest you must usually request from the producer and its not often they have it either.

The most important variations are SG, viscosity with at least two temperatures and vapor pressure for transportation analysis. For heavies, you'll need the heat transfer data, but most of the time, I just have to assume something reasonable. The producers just don't know much more in many cases.

The only times I have ever had to use compositional models was for a very large gas gathering system connecting many fields with high CO2 and up to 40% N2 content and or fields H2S in the gas and they had to be very carefully blended with better streams before putting it into the main interstate transmission line. Still we ignored the condensates.

http://virtualpipeline.spaces.msn.com

"What gets us into trouble is not what we don't know, its what we know for sure" - Mark Twain

 

[ColourfulFigsnDiags]

I hear what you are saying BigInch, but I have just come off comparing a black oil model and a compositional model in some well simulations, and the difference can be quite large.

In this instance the thermodynamic differences between the two models were negligible, but the hydraulic differences were, in some cases, huge, with WHFPs differing by upwards of 40 bar between the two models. Since most of the pressure drop in wells is due to the hydrostatic component, in this case the difference between the two models was the phase envelope. This is not a particullarly heavy oil, and we did the same analysis for some of our gas producers, and even then the results were generally less different, but still significant for us to abandon the black oil model data and rely on the compositional model data obtained via crude assays and other reservoir samples.

Read the Eng-Tips Site Policies at FAQ731-376

 

[BigInch]

That's why I said, unless you have reason to do it. A deep well could certainly qualify whereas a pipeline transport case would most likely not, but there are also qualifiers that go with that too.

Its really just one of my biggest pet peeves, building models way too detailed for the answers needed. Didn't mean to make an issue out of it. I'll shut up now.

http://virtualpipeline.spaces.msn.com

"What gets us into trouble is not what we don't know, its what we know for sure" - Mark Twain