Petroleum Engineering Department

PETROLEUM PROPERTIES
Lecture 5

Interconversion of Various Distillation Data

By: Foad Yousefi

Average boiling point

Application and estimation of various boiling points are discussed by Van Winkle. Since the
materials boil over a range of temperature, any one average boiling point fails to be useful for
correlation of all properties. The most useful type of ABP is MeABP, which is recommended for
correlation of most physical properties as well as calculation of Watson Kw. However, for
calculation of specific heat, VABP is recommended. In Example MeABP is 121.7 C which varies
from 126.7 for the ASTM D 86 temperature at 50 vol% distilled (T50). However, based on our
experience, for narrow boiling range fractions with SL < 0.8 C/% the MeABP is very close to 50%
ASTM temperature.

As an example, for a gas oil sample with ASTM temperatures of 261.7, 270, 279.4, 289.4, and
307.2 C at 10, 30, 50, 70, and 90 vol%, the MeABP is calculated as 279, which is very close to
50% ASTM temperature of 279.4 C For this fraction the value of SL is 0.57 C which indicates the
boiling range is quite narrow. Since none of the average boiling points defined here represent the
true boiling point of a fraction, the 50% ASTM temperature may be used as a characteristic boiling
point instead of average boiling point. In this case it is assumed that the difference between these
temperatures is within the range of experimental uncertainty for the reported distillation data as
well as the correlation used to estimate a physical property.

Interconversion of Various Distillation Data

Work to develop empirical methods for converting ASTM distillations to TBP and EFV
distillations began in the late 1920 and continued through the 1950 and 1960 by a large number of
researchers. All of the correlations were based on discordant experimental data from the literature.
Experimental ASTM, TBP, and EFV data on which the empirical correlations are based suffer a
lack of reproducibility because there were no standardized procedures or apparatus available. All
of these correlations were evaluated and compared to each other by House et al.to select most
appropriate method for inclusion in the API-TDB. As a result of their evaluations, the following
methods were adopted in the API Data Book as the best method:

• Edmister-Pollock for ASTM to TBP

• Edmister-Okamoto for ASTM to EFV

1
By: Foad Yosefi
 Petroleum Properties Lecture 5

• Maxwell For conversion of TBP to EFV.

Most of these correlations were in graphical forms and inconvenient for computer applications.

Riazi-Daubert Method

Riazi and Daubert methods for the interconversion of various distillation data are based on the
generalized correlation for property estimation of hydrocarbons. Available distillation temperature
and specific gravity of the fraction are used as the input parameters to estimate the desired
distillation data in the following form:

𝑻𝒊 (𝒅𝒆𝒔𝒊𝒓𝒆𝒅) = 𝒂 [𝑻𝒊 (𝒂𝒗𝒂𝒊𝒍𝒂𝒃𝒍𝒆)]𝒃 𝑺𝑮𝒄 1

where Ti (available) is the available distillation temperature at a specific vol% distilled and
Ti(desired) is the desired distillation data for the same vol% distilled, both are in kelvin. SG is the
specific gravity of fraction at 15.5 C and a, b, and c are correlation parameters specific for each
conversion type and each vol% point on the distillation curve. For example, if this equation is used
to convert ASTM to EFV at 10%, Ti (available) is ASTM temperature at 10% and Ti(desired) is
the EFV temperature at 10% and constants a, b, and c are specific for this conversion type at 10%
of volume vaporized.

ASTM D 86 and TBP Conversion

If distillation data available are in the form of ASTM D 86 and desired distillation is TBP, Eq. (1)
can be used, but for this particular type of conversion value of constant c for all points is zero and
the equation reduces to:

𝑻𝑩𝑷 = 𝒂 [𝑨𝑺𝑻𝑴 𝑫 𝟖𝟔]𝒃 2

where both TBP and ASTM temperatures are for the same vol% distilled and are in kelvin.
Constants a and b at various points along the distillation curve with the range of application are
given in Table 1.

For a total of 559 data points for 80 different samples, Eq. (2) gives an average absolute deviation
(AAD) of about 5 C while the Edmister-Pollock method gives an AAD of about 7 C.

2
By: Foad Yosefi
 Petroleum Properties Lecture 5

Generally predictions at 0% give higher errors and are less reliable. Equation (2) can be easily
reversed to predict ASTM from TBP data, but this is a rare application as usually ASTM data are
available.

TABLE 1-Correlation constants for Eq. (2).

If TBP distillation curve is available then ASTM curve can be estimated as:
𝟏
𝟏 𝒃 𝟏
𝑨𝑺𝑻𝑴 𝑫 𝟖𝟔 = ( ) (𝑻𝑩𝑷)𝒃 3
𝒂
where constants a and b are given in Table 1 as for Eq. (2).

ASTM D 86 and EFV Conversions

Application of Eq. (1) to this type of conversion gives:

𝑬𝑭𝑽 = 𝒂 (𝑨𝑺𝑻𝑴 𝑫 𝟖𝟔)𝒃 (𝑺𝑮)𝒄 4

where constants a, b, and c were obtained from more than 300 data points and are given in Table
2. Equation (4) was evaluated with more than 300 data points from 43 different samples and
gave AAD of 6 C while the method of Edmister- Okamoto gave an AAD of 10 C.

3
By: Foad Yosefi
 Petroleum Properties Lecture 5

TABLE 2-Correlation constants for Eq. (3).

In using these equations if specific gravity of a fraction is not available, it may be estimated from
available distillation curves at 10 and 50% points as given by the following equation:

where constants a, b, and c for the three types of distillation data, namely, ASTM D 86, TBP, and
EFV, are given in Table 3. Temperatures at 10 and 50% are both in kelvin.

TABLE 3---Correlation constants for Eq. (4).

4
By: Foad Yosefi
 Petroleum Properties Lecture 5

SD to ASTM D 86 Conversions

The equation derived from Eq. (1) for the conversion of simulated distillation (ASTM D 2887) to
ASTM D 86 distillation curve has the following form:

𝑨𝑺𝑻𝑴 𝑫 𝟖𝟔 = 𝒂 (𝑺𝑫)𝒃 (𝑭)𝒄 6

where constant F is a parameter specifically used for this type of conversion and is given by the
following equation:

𝑭 = 𝟎. 𝟎𝟏𝟒𝟏𝟏 (𝑺𝑫 𝟏𝟎%)𝟎.𝟎𝟓𝟒𝟑𝟒 (𝑺𝑫 𝟓𝟎%)𝟎.𝟔𝟏𝟒𝟕 7

in which SD 10% and SD 50% are the SD temperatures in kelvin at 10 and 50 wt% distilled,
respectively. Parameter F calculated from Eq. (7) must be substituted in Eq. (6) to estimate ASTM
D 86 temperature at corresponding percent point expressed in volume basis. Equation (6) cannot
be used in a reverse form to predict SD from ASTM D 86, but this type of conversion is usually
not desired as most predictive methods use ASTM D 86 data while laboratories report SD data.
Constants a, b, and c in Eq. (6) were obtained from 81 different samples and 567 data points and
are given in Table 4 with the range of SD data at each percentage along the distillation curve.

TABLE 4---Correlation constants for Eq. (6).

The procedures given in this section should be used with the range of data specified in Tables 1-
4. Use of these equations outside the specified ranges could cause large errors.

5
By: Foad Yosefi
 Petroleum Properties Lecture 5

Example 5.1.

For a blend of naphtha-kerosene sample, ASTM, TBP, and EFV distillation curves are given in
the APITDB. These data are represented in Table 5. Use the Riazi-Daubert methods to predict
EFV curve from TBP curve.

TABLE 5.Data on various distillation curves for a naphtha-kerosene blend.

6
By: Foad Yosefi