8 replies to this topic

[daryon]

Hi all,

(sorry this is a long post, please bear with me)

I have question i've been trying to figure out the answer to, and hoped someone might be able shed some light on it. Its regarding crude oil characterisation. I've been working in the oil and gas industry for a few years now and have a bit of experience with crude oil compositions that are typically provided by the upstream reservior guys to the process engineers for detailed design of processing facilities.

A typical crude oil contains say 40% alkanes, 40% Cycloalkanes, 15% Aromatics, and 5% asphaltenes, the numbers aren't important here, but these four classes of hydrocarbons make up the bulk of all crude oils. A paraffin based crude will be rich in higher molecular weight alkanes and lean in asphaltenes.

What i don't understand is why most crude oil compositions (that i've seen) don't list any components apart from alkanes. Or list a few cycloalkanes and aromatics components with small mol fractions, but no where the factions that make up a typical crude. Take the below composition for example; this was provided by a client for a medium crude oil (32 API) to be produced from the gulf of mexico.



COMPONENTS Mole %
CO2 0.45
H2S 0
Nitrogen 0.47
Methane (C1) 20.03
Ethane (C2) 3.73
Propane (C3) 5.1
i-Butane (C4) 0.89
n-Butane (C4) 2.75
i-Pentane (C5) 1.32
n-Pentane (C5) 1.85
n-Hexane (C6) 3.13
Mcyclopentane (C5) 0.51
Benzene (C6) 0.11
Cyclohexane (C6) 0.22
n-Heptane (C7) 2.87
Mcyclohexane (C7) 0.42
Toluene (C7) 0.29
n-Octane (C8) 2.94
E-Benzene (C8) 0.24
m&p-Xylene (C8) 0.11
o-Xylene (C8) 0.17
n-Nonane (C9) 2.7
n-Decane (C10) 2.91
n-C11 2.64
n-C12 2.17
n-C13 2.14
n-C14 1.94
n-C15 2.02
n-C16 2.8
n-C17 1.76
n-C18 2.07
n-C19 1.38
n-C20 1.14
n-C21 1.60
n-C22 0.95
n-C23 1.05
n-C24 0.82
n-C25 0.76
n-C26 0.72
n-C27 0.67
n-C28 0.74
n-C29 0.59
C30+* 18.84
Total 100.01


---Hypothetical---
C30+ Properties
MW 750
Density @ 60F 1.01 g/cm3

Here we have a couple of cycloalkanes and aromatics for C6, C7,C8 but the total fraction of cycloalkanes and aromatics is less than 0.02. I can't believe this is an accurate representation of the actual crude composition.

The FPSO project have recently finished the process simulation for had a crude oil composition which just contained alkanes. THis I find even harder to believe. The component mol fractions we given up to C20 as alkanes and two hypothetical groups (C20+ and C30+) were provided.

Why are crude oil compositions often given with just the alkane components? Is this to do with accuracy of the composition and it being pointless to indentify the compostion of crude oil in the detail of determing what fraction of C8 hydrocarbons are alkanes, cycloalkanes and aromatics. Or is this labs being lazy. Cycloheptane has significantly different properties to heptane and toluene.


C7 components:


Cyclo-heptane
MW =98.18
SG = 0.810
BP (oC) =118 -120

n-heptane
MW = 100.2
SG = 0.684
BP (oC) = 98.4

toluene
MW = 92.13
SG = 0.866
BP (oC) = 110.8


I guess if the component fractions are identified by TBP distillation then the lab guys don't have a clue whats actually in the fraction they just know the boiling point and then say well thats C7, and it goes in the composition list as heptane?

I could understand what was going on if the composition specified C7 cut and gave boiling point, density and molecular weight, but it is given just as C7 which we take for pure heptane. If there is a significant amount of toluene (which is likely to boil in the same fraction as octane won't boil till 126 degC) then the density of this fraction could potentially be way off that of pure heptane.

Any light you can shed on this would be great. My understanding of crude characterisation is certainly lacking here

Thanks in advance

Daryon

[gvdlans]

I would say there are two main reasons:

1) Crude oil is a mixture of countless different components, especially when it comes to heavier hydrocarbons. It is not so easy (and therefore expensive) to identify all the different components...

2) For most purposes (e.g. equipment design, refinery operations, economic/commercial calculations), there is no need to have a more detailed composition. It is sufficient to know characteristics such as:

- BS&W (Basic Sediment and Water)
- sulphur content
- ASTM boiling curve

See http://www.petrotech.../crudechar.html for the typical information in a crude assay.

[chemtan]

You can have more detailed composition of crude oil... there is no doubt in it.

But the question comes, do you need it? I think it is because there might be no benefit of knowing those components when your downstream plant (refinery) is working to produce\process alkanes mostly.

BTW, just looking at comp. that you gave, it already reached 100%, so wondering if there is anything else there uncovered in that crude now.

[daryon]

Thanks for your replies.

I guess the most import thing is that the characterised crude oil gives a fairly accurate representation of how the actual crude oil will behave as it passes through the process. It therefore doesn't really matter what components are entered into the simualtion software as long as the characterised crude oil exhibts the correct PVT behavoir and physical properties when the component flash and physical property calculations are performed.

The most important aspect is that the characterised crude oil matchs laboratory data when property checks are performed like GOR, viscosity, density, etc. If you get bad prediction of properties and PVT behavoir you could potentially end up with poorly designed process.

Interestingly on a previous project, we were given a crude oil compostion C1 to C10, then a couple of hypothetical components. When we put this composition into HYSYS and performed property checks the HYSYS soultion GOR was way above the laboratory reported data. After discussions with the client, and really against our will we ended up removing a signicant amount of gas from the simualted well stream upstream of the process (in order to match reported GOR). I don't know how the process actually faired but I wouldn't be surpised if the gas system is undersized and restricts production from the process!

[JoeWong]

Interestingly on a previous project, we were given a crude oil compostion C1 to C10, then a couple of hypothetical components. When we put this composition into HYSYS and performed property checks the HYSYS soultion GOR was way above the laboratory reported data. After discussions with the client, and really against our will we ended up removing a signicant amount of gas from the simualted well stream upstream of the process (in order to match reported GOR). I don't know how the process actually faired but I wouldn't be surpised if the gas system is undersized and restricts production from the process!

I am not sure if i understood your statement correctly. Are you key in the "crude composition" and flash it at specified condition and obtain the GOR from the flashing ?

Normally the composition given is dry basis at particular condition. You may need to obtain the vapor composition and liquid composition, adjust the vapor & flow till the GOR matching lab reported GOR.

[daryon]

I am not sure if i understood your statement correctly. Are you key in the "crude composition" and flash it at specified condition and obtain the GOR from the flashing ?

Normally the composition given is dry basis at particular condition. You may need to obtain the vapor composition and liquid composition, adjust the vapor & flow till the GOR matching lab reported GOR.
[/quote]


Hi Joe,

The reservoir fluid sample (which was a bottomhole sample) was given on a dry basis, after entering the composition into HYSYS (including hypo's) we saturated with water at reservoir conditions (137 barg, 126 degC). We then flashed the stream to conditions which the laboratory perfomed their flash test (13.8 barg and 38 degC) which was the assumed separator conditions.

We compared the HYSYS GOR with the laboratory reported GOR and they were way off. Lab suggesting 500 Scf/bbl and HYSYS predicting 850 Scf/bbl. The lab perform a single stage flash which is what we modelled in HYSYS, so the GOR should be in fairly close agreement if the reservoir sample is characterised properly?

I agree with your appoarch of using a gas adjust stream to match reported GORs and this is typically what I would do. This is the only occasion which I have had a higher GOR predicted by HYSYS and had to remove gas in order to match the reported GOR. This is what makes me doubt the accuracy of the characterised crude oil composition we were given.

[JoeWong]

Daryon,
The way you described in last paragraph pretty inline with i would do for most of the cases...
The logic that you presented in first paragraph should have missing something (i think)...

(i) Normally there is excess produced water in the separator. It may give some impact to your fluid characterization. It may not only saturate at reservoir condition but may be saturate at separator condition.

(ii) I do think that there is some mis-interpretation of the reported figure. Is the given "composition" at separator condition, reservoir or... Is "sample" came from existing well...

You can imagine. Information pass from Client Reservoir engineer to lab. specialist, then from lab specialist to client process engineer, and then to contractor/ consulting engineer...There may be communication break-down here. I have a lot of this type of bad experiences. Really suggest you discuss and firm-up this CRITICAL basis with client (at the beginning of the job).

[daryon]

Hi Joe,

(i) The crude oil was saturated at reservoir conditions, the higher T&P meant it will hold more water. When the T&P is dropped to separator conditions there is a separate aqueous phase. Therefore i don't think additional water will affect the composition.

(ii) The sample taken subsurface (bottomhole) it is a reservoir fluid sample.

I think you're right about the confusion and mis-interpretation somewhere along the line. We did our best to express our concerns to the client who were frankly much more concerned with project schedule than discussions on firming up the design basis.

I still have in the back of mind the possibility that the characterisation work / testing perfomed by the lab was poor, I appreciate you can't comment without seeing a reservoir fluid study.

[Zauberberg]

Daryon,

It all depends on what do you need characterization for - that will define the level of accuracy required for feed definition.

For example, if you are looking for sizing a production separator, a minimum level of details is required. You just have to obtain composition up to C5 or C6, and to lump other components into C6+ or C7+ groups. However, if you need to design the entire plant consisting of multistage separators, compressors, fractionation towers etc, then accurate feed characterization is a must.

I remember doing the same thing for gas plant design in Oman, and I wasn't very much happy with the form of PVT report (not too many data that you can actually use for characterization). It was a 3-weeks task, and it required a lot of efforts and fine tuning within the Oil Manager environment.