Precision and Equivalence of Automatic and Manual Flash Point Apparatus

Abstract

Abstract The flash point of a liquid material is used to classify whether the material is flammable or combustible. Historically, flash point determination has been done using manual flash point apparatus. However, the advent of automatic flash point apparatus, especially in the petroleum and related industries, brought about improved testing efficiency such that manual flash point determination is becoming the less preferred way of determining flash point. Three ASTM standard test methods, namely, D 56 “Flash Point by Tag Closed Cup Tester,” D 92 “Flash and Fire Points by Cleveland Open Cup,” and D 93 “Flash Point by Pensky-Martens Closed Cup Testers” are the most common test methods used in the various specifications and regulatory requirements for determining the flash point of materials. Three interlaboratory studies were conducted in 1991 to determine new precision values for the ASTM D 56, D 92, and D 93 flash point test methods, and to determine if the flash points obtained by the automatic apparatus are statistically equivalent to the flash points obtained by the manual apparatus in each of these three standard test methods. The pooled repeatability at a 95% confidence level for the D 56 flash point was found to be 1.1°C (manual) and 1.2°C (automatic) for flash points below 60°C; and 1.4°C (manual) and 1.6°C (automatic) for flash points at or above 60°C. The pooled reproducibility for D 56 at a 95% confidence level was found to be 4.1°C (manual) and 4.3°C (automatic) for flash points below 60°C; and 5.8°C (manual) and 5.1°C (automatic) for flash points at or above 60°C. For D 92, a repeatability at a 95% confidence level of 4°C (manual) and 8°C (automatic); and a reproducibility at a 95% confidence level of 13°C (manual) and 18°C (automatic), were obtained. The precision for D 93 was shown to be dependent on the flash point value, with a repeatability at a 95% confidence level of 0.032 X °C (manual) and 0.035 X °C (automatic), and a reproducibility at a 95% confidence level of 0.073 X °C (manual) and 0.078 X °C (automatic). Standard statistical evaluation tests were performed on the data for the three test methods. Essentially, no statistically significant differences were found between automatic and manual D 56, automatic and manual D 92, and automatic and manual D 93 flash point results. Thus, the interlaboratory studies of the three flash point test methods indicated that automatic flash point results are equivalent to manual flash point results.