Friday, November 28, 2014

Repost of our most popular question in 2014: What are the Flash Point Calculation Methods?

This is a question commonly asked of Technical Support and the most frequently viewed post in 2014 - so i figured we would post the answer here to help those that don't want to contact Technical Support. 

Flash point - The minimum temperature at which a liquid gives off a vapor in sufficient concentration to ignite when tested by the following methods:

1.Tagliabue Closed Tester (See American National Standard Method of Test for Flash Point by Tag Closed Tester, Z11.24 1979 [ASTM D56-79]).
2.Pensky-Martens Closed Tester (See American National Standard Method of Test for Flash Point by Pensky-Martens Closed Tester, Z11.7-1979 [ASTM D93-79]).
3.Setaflash Closed Tester (See American National Standard Method for Test for Flash Point by Setaflash Closed Tester [ASTM D 3278-78]).
Pensky-Martens (ASTM D93) is used for the flash point of fuel oils as well as viscous materials and suspension of solids.

Tagliabue (ASTM D56) is used for the determination of the flash point of all mobile liquids flashing below 200F except products classed as fuel oil. The Tag method is recommended for liquids having viscosities less than 5.5 cSt at 40C that do not contain suspended solids or do not have a tendency to form a surface film while under test.
(1003 12/15 TLM:  Added D86 to the API method below)
(1312 12/15 TLM:  Added the edition and year to API Data Book below)
(1328 12/15 TLM:  Changed the spelling of Pensky-Martens below)
In PRO/II and in ROMeo:
Flash point can be calculated by the Nelson method (FLPOINT=NELSON) or by the Pensky-Martens method (FLPOINT=API).  Nelson is the default.
PRO/II keyword syntax:
NELSON is default.
The other is API 2B7.1(Penske-Martins Closed Cup) .
The Nelson method is described in "Petroleum Refinery Engineering" by W.L. Nelson, 4th edition, McGraw-Hill Book Company, p 131-133. The reference does not mention any restrictions on the range of applicability. I have copies of articles by Nelson in the 1944 Oil and Gas Journal in which he states his correlation is "probably not accurate enough to differentiate between diferent kinds of flash points".
For the Nelson method, set T= (IP + 2 x 5%pt + 10%pt)/ 4 and then the FP= 0.64 x T - 100. Note that T is an average T that may be defined differently by other implementations of the Nelson flash point distilled fraction equation. The temperature is in F and the percentages are volume percents of the ASTM D86 distillation. Since the initial point for simulated refinery streams is notoriously inaccurate, I am not surprised that inaccuracies are found in the Nelson method simulated results.
The API Procedure 2B7.1 is found in the API Data Book, 5th ed., 1992 (API 2B7.1). This is the Pensky-Martens closed cup method. The recommended range of applicability is between -15F to 325F flash points.
For the API method 1/FP = -0.014568 + 2.84947/T1 + 1.903E-03 ln(T1) where T1= ASTM D86 10% temperature for petroleum fractions or normal boiling point for pure compounds in degrees Rankine and FP is the flash point in degrees Rankine. There is a Figure 2B7.2 in the API Data Book that represents the equation written above.
A comment on the correlation of plant data to predicted results...
The best any simulator can do is to "simulate" a lab procedure. ASTM Lab tests are quite specific as to time and temperature handling of samples. Errors can creep into the results in taking and storing the sample as well as the actual lab analysis.  I do not know how close the simulated results are to plant data, but for the D93 lab test itself, answers for duplicate samples analyzed at different labs are acceptable within 6F.

As always, feel free to reach out to your local technical support team.