PETROVIETNAM

QUALITY ASSESSMENT OF HYDROCARBON PVT DATA

Nguyen Van Do, Pham Truong Giang, Hoang Long, Le Thi Thu Huong
Vietnam Petroleum Institute
Email: donv@[Link]

Summary
PVT data are essential for proper field development plan, accurate reserve estimation, reservoir simulation, material
balance calculation, and optimum fluid recovery plans. High quality and accurate PVT data can, therefore, reduce uncertainty
in reservoir fluid properties and set the stage for reservoir engineering modelling whilst improving the technical work on which
investment decisions are made. Poor quality data will, however, result in lost time due to rework and additional studies, poor
development plans, and inefficient investment.
This paper focuses on the testing of the quality of the PVT data from DH-8X-DST#4 performed in the lab based on Hoffman
plot, material balance, and Buckley plot.
Key words: PVT, material balance, Hoffman, Buckley, Dai Hung field.

1. Introduction Where:

Quality control (QC) of PVT data is very important. Pc: Critical pressure (psia)
Improper quality control or any error in the recorded data T: Separator temperature (oR)
can be misleading in the next steps of fluid modelling,
fluid properties measurement, and construction of Tb: Boiling temperature (oR)
thermodynamic model for representative reservoir fluid. Tc: Critical temperature (oR)
It is, therefore, important to conduct fluid sampling and
K: Equilibrium ratio
recombination using standard protocols to minimise
errors and uncertainties. In this paper, we introduce quality A plot of Log (K) versus F for each component of the
control of PVT measurements on collected gas and liquid separator liquid and reservoir fluid compositions will
samples within a reservoir to select the representative be expected to yield a linear graph which implies high
reservoir fluid. After that, the representative reservoir fluid quality data where all points plotted are expected to fall
will be used for simulation. close to the line of best-fit.

2. Methodology Equilibrium ratios trend (K): The equilibrium ratio

[2], Ki, of a given component in a hydrocarbon mixture is
a) Graphical technique defined as the ratio of the mole fraction of the component
Graphical plots [1] of the non-compositional data as a in the gas phase, yi, to that of the mole fraction of the
function of pressure were made. The following properties component in the liquid phase, xi. Mathematically, the
were plotted against pressure to check the quality of the relationship is expressed as:
PVT data: relative volume, Y-function, density, solution
GOR and oil FVF, the graphs are expected to show a
smooth trend.
The K-value measures the tendency of the component
b) Hoffman plot
to remain or escape to the gas phase. It is essentially a
The Hoffman plot [1] is another method of evaluating property that measures the volatility of the component
the quality of compositional data. This method utilises a at a pressure and temperature. This tendency suggests
log-linear plot of K-value versus Hoffman factor, F. F (psi/ that in a multicomponent system, components’ Ki-value
o
R) is defined as follows: should follow the trend of their decreasing K-values with
the increasing order of their normal boiling point; that is,
F = [(LOG(Pc) - LOG(14.7)] × 1 1 1 1
Tb T Tb Tc KN > KCl > KCO > KC > KH S > KC > …
2 2 2 2 3

c) Material balance
Date of receipt: 23/5/2017. Date of review and editing: 23/5 - 26/5/2017. Date of approval: 26/5/2017.

PETROVIETNAM - JOURNAL VOL 6/2017 35

PETROLEUM EXPLORATION & PRODUCTION

The material balance test [1] was used

Buckley plot
to evaluate compositional consistency. It 3
was applied on the separator liquid and HC component
2 Non-HC component
reservoir fluid compositions. The material
1 Linear (HC component)
balance equation is expressed below.

Log K
Fzi = Lxi + Vyi 0
0 2 4 6 8 10
Where: -1
y = -0.5666x + 3.8701
-2
F: Total moles of feed R² = 0.8813
-3
L: Total moles of separator liquid
V: Total moles of separator vapour
Figure 1. Buckley plot [4]
zi: Mole fraction of component i in the
Table 1. Fluid data [5]
feed
xi: Mole fraction of component i in the No. Fluid data
liquid 1 Sample name Dai Hung-8X-DST#4
2 Sample description Surface sample
yi: Mole fraction of component i in the
3 GOR 1103.3 scf/bbl
vapour
4 Reservoir pressure 4,050psia
This equation can be rearranged to the 5 Reservoir temperature 245oF
following: 6 API 37
yi/zi = F/V – (L/V)(xi/zi) 7 Oil volume factor (Bo) 1.65

A plot of yi/zi against xi/zi should give Table 2. Separator and recombined reservoir fluid data [5]
a straight line with a slope of –L/V and %mole
Component
an intercept of F/V. Any deviations from Separator gas Separator liquid Reservoir fluid
a straight line reflect material balance N2 0 0 0
discrepancies. Reservoir fluid compositions CO2 3.611 0.713 2.416
are typically calculated from a mathematical H2 S 0 0 0
CH4 77.034 8.293 48.684
recombination of flashed vapour and liquid
C2H6 9.55 4.073 7.291
compositions. When feed compositions
C3H8 6.355 7.213 6.709
are determined from a mathematical i-C4H10 1.291 3.074 2.026
recombination of the separator vapour and n-C4H10 1.375 4.436 2.637
separator liquid compositions, a perfectly i-C5H12 0.377 2.521 1.261
straight line should be observed in the n-C5H12 0.211 1.901 0.908
plot. In this case, the plot is sometimes C6 0.14 4.998 2.144
useful for identifying discrepancies in the C7+ 0.056 62.778 52.8
reported compositions. The reciprocal of Sum 100 100 100

the slope of the line can then be used to

calculate the GOR and compare that with d) Buckley plot
the reported GOR. The liquid density and The Buckley plot (Figure 1) is expressed by the semi logarithmic
molecular weight must be provided so plot of K-value versus the square of the critical temperature Tc of the
that the conversion from moles to barrels reservoir fluid [4]. It gives a straight line with negative slope for the
can be made. When the feed composition light hydrocarbon components and any significant deviation indicates
is truly independent from the vapour and possible non-equilibrium separation or numerical errors in the data
liquid compositions, the material balance reporting. As a component becomes less paraffinic then the deviation
test provides a good check on the overall from the linearity becomes greater.
consistency of the analyses [3].

36 PETROVIETNAM - JOURNAL VOL 6/2017

 PETROVIETNAM

Table 3. Differential liberation test at 245oF [5]

Pressure Oil FVF, Bo GOR Oil density Gas FVF Gas density
psia vol/svol scf/bbl lb/cf vol/svol g/cc
5,000 1.602 1103.3 40.5
4,673 1.610 1103.3 40.3
4,359 1.618 1103.3 40.1
4,050 1.627 1103.3 39.9
3,580 1.641 1103.3 39.6
3,255 1.651 1103.3 39.3
2,700 1.559 862.7 40.1 0.00648 0.726
2,100 1.471 655.0 41.1 0.00834 0.732
1,500 1.392 473.4 42.1 0.01176 0.745
1,000 1.323 329.5 43.2 0.01807 0.766
525 1.254 193.8 44.3 0.03535 0.841
145 1.18 74.7 45.7 0.13137 1.013
15 1.099 0 47.8 1.33064 1.45

3. QC of DH-8X-DST#4 Sand 0
10
9 The result of PVT analysis from DH-8X-
DST#4 is shown in Table 1 - 3.
8
7 Based on Table 2, the equilibrium
ratio for each component was calculated
6
as shown in Figure 2, which indicates a
5
consistency in the laboratory-reported
4 compositional analysis.
3
Figure 3 is the resulting material
2 balance plot when yi/zi is plotted against
1 xi/zi. In this case, the plot from the material
0 balance of separator liquid and reservoir
4 2 6 8 10 10 12 12 6 fluid compositional data shows a linear
2 3 4 4 5 5
trend. This confirms the high quality and
accuracy of compositional data obtained.
Figure 2. K-value The slope of the line in the figure is -1.4248,
which can be converted to a GOR of 1,102.5
3 scf/bbl by taking the negative reciprocal of
the slope and converting it to the measured
2.5 unit. This calculated GOR is about 0.01%
lower than the reported laboratory value
2
y = -1.4248x + 2.4248 of 1,103.3 scf/bbl. The difference is very
R² = 1 small, indicating that the liquid molecular
y z

1.5
weight, liquid density, and recombination
1 ratio all appear reasonably consistent with
the reported GOR.
0.5
The Hoffman plot is a standard
0 technique for evaluating the consistency
0 0.5 1 1.5 2 2.5 3 of the recombined fluid samples through
z x
a graphical technique. Usually the plus
fraction (e.g. heptane-plus (C7+)) is not
Figure 3. Material balance plot

PETROVIETNAM - JOURNAL VOL 6/2017 37

PETROLEUM EXPLORATION & PRODUCTION

been applied for a long time, but they are still

3
C1
very useful today. These methods are material
2.5
balance, Hoffman and Buckley plots. The
2 C2 material balance plot can be used as a rigorous
1.5 CO2
criterion for consistency while the Hoffman plot
1
Log K

iC4 C3 Series1 and the Buckley plot are considered qualitative

0.5
nC4 Linear (Series1) assessment of data quality.
iC5 0
-2 -1.5 -1 -0.5 0 0.5 1 1.5
nC5 -0.5 Application of this methodology on the PVT
C6 -1 data from DH-8X-DST#4 under study showed
-1.5 that the samples can be used as representative
Hoffman factor
for the whole reservoir.
Figure 4. Hoffman plot
References

10 1. Nnaemeka Nnamdi Nnabuo.

C1 9 y = - 2.9404x + 4.2056 Interpretation of laboratory PVT analysis result (a
8 R² = 0.9913
C3 case study of a Niger delta field). SPE 172412-MS.
C2 7
SPE Nigeria Annual International Conference
6
C4 and Exhibition, Lagos, Nigeria. 5 - 7 August,
Log K

5 C5
iC4 Series1 2014.
4
iC5 Linear (Series1)
3 2. Tarek Ahmed. Equations of state and PVT
2
C6 analysis. Applications for Improved Reservoir
1
Modelling. 2016.
0
-2 -1 0 1 2 3. John J. Lawrence. Quality assessment and
5
consistency evaluation of hydrocarbon PVT data.
Figure 5. Buckley plot SPE 13784-MS. SPE International Petroleum
Technology Conference held in Doha, Qatar. 7
included in Hoffman consideration because it is generally not - 9 December, 2009.
measured accurately [2]. To further control the quality of the
4. Julius Akpabio, Sunday [Link],
recombined compositions listed in Table 2, the compositions were
Oluwatoyin Akinsete. PVT fluid sampling,
used to generate the Hoffman plot as shown in Figure 4, which
characterization and gas condensate reservoir
confirms that the separator liquid and reservoir compositional data
modeling. Advances in Research. 2015; 5(5):
are of high quality, as the plotted points lie close to the line of best
p. 1 - 11.
fit for components C1 through C6 for all pressure.
5. PVEP POC. PVT report of DH-8X-DST#4.
The plot from Buckley plot (Figure 5) is expressed by the
relationship of Log K versus Tc for the reservoir fluid. A semi log plot of 6. John [Link]. Getting the best out of
K value versus the square of the critical temperature gives a straight fluid samples. SPE 29227-PA. 1994; 46(9).
line with a negative slope for the hydrocarbon components (C1 - C6). 7. Alan Graham Stephen, David [Link],
This confirms the high quality and accuracy of the compositional Tim Dodd, Wayne Kriel. PVT data quality: Round
data obtained. Robin result. SPE 116162-MS. SPE Annual
4. Conclusions Technical Conference and Exhibition. Denver,
Colorado, USA. 21 - 24 September, 2008.
In order to obtain trustworthy PVT fluid properties, the quality of
the samples must be examined carefully. Three other techniques for
evaluating the consistency of flash and composition data have also

38 PETROVIETNAM - JOURNAL VOL 6/2017