Flash Point Correlation API 2B7.1 (actual)

[Fuel]

Hello, colleagues.
In the SIMSCI PRO-II I can select one of two correlations: API 2B7.1 (1987) or API 2B7.1 (2007). These correlations provide very different results (up to 10°C for typical diesel distillation curve). It seems to me that the new method gives a too low value. I don't have the formula used to calculate flash point according to the API 2B7.1 (2007).
Does anyone known this formula? I want to understand why there is such a difference between the results of the old and new methods.
Thanks.

[pmx]

Hi Fuel.

PRO/II usually lets you select different correlations ("procedures") from API Technical Data Book (TDB).

I am not familiar with API Technical Data Book (TDB) procedure 2B7.1 from 2007. The latest flash point procedure I am aware of, is from 1999. Find below the equations for the 1987 and 1999 versions:

API TDB Procedure 2B7.1 (1987):
1/FP = -0.014568 + 2.84947/T1 + 1.903E-03 ln(T1)
where
T1: ASTM D86 10% temperature for petroleum fraction in degrees Rankine
FP: flash point in degrees Rankine
This FP was correlated to closed cup method values.
Validity range: FP between -15ºF to 325ºF.

API TDB Procedure 2B7.1 (1999):
Closed cup: FP = 0.69*T1 - 118.2 (equation 2B7.1-1)
Open cup: FP = 0.68*T1 - 109.6 (equation 2B7.1-2)
where
T1: ASTM D86 10% temperature for petroleum fraction in degrees fahrenheit
FP: flash point in degrees fahrenheit
Validity range: FP between 0ºF to 450ºF.

You can check whether the results PRO/II gives for API 2B7.1 (2007) are the same as those obtained with API TDB Procedure 2B7.1 (1999).

Please note that the differences between both results for diesels you state, of up to 10°C (18ºF) is not that unusual:
1) The ASTM D93 lab test itself, answers for duplicate samples analyzed at different labs are acceptable within 6ºF.
2) Equation 2B7.1-1 (1999) has an average absolute deviation of 9,6ºF.
3) Due to these kind of differences, Gerald Kaes (in his book Refinery Process Modeling) recommends that "a predicted flash point within plus or minus 10 to 15ºF (5.6 to 8.3ºC) to the laboratory test value is reasonable.

Hope this helps. If you get the equations behind PRO/II API 2B7.1 (2007) flash point calculations, please post them here.