ASTM D2699 test method for Research Octane Number of Spark-Ignition Engine Fuel
18. Precision and Bias
18.1 Bracketing - Equilibrium Fuel Level Procedure A and C.R. Procedure C:
18.1.1 90.0 to 100.0 Research O.N. Range - The precision of this test method for Research O.N. values between 90.0 and 100.0 based on statistical examination of interlaboratory test results by the bracketing–equilibrium fuel level or C.R. procedures is as follows:
18.1.1.1 Repeatability - The difference between two test results, obtained on identical test samples under repeatability conditions would, in the long run, in the normal and correct operation of the test method, exceed 0.2 O.N. only in one case in twenty.
18.1.1.2 Reproducibility - The difference between two single and independent results obtained on identical test samples under reproducibility conditions would, in the long run, in the normal and correct operation of the test method, exceed 0.7 O.N. only in one case in twenty.
18.1.1.3 The above repeatability is based on the replicate O.N. results obtained by the ASTM Motor National Exchange Group (NEG) participating in cooperative testing programs from 1983 through 1987 and 1994. Between 90 and 100 Research O.N., the repeatability standard deviation is 0.08 unaffected by octane level. This average standard deviation has been multiplied by 2.772 to obtain the limit value.
18.1.1.4 The above reproducibility is based on the combined NEG monthly sample testing program data from 1988 through 1994, the Institute of Petroleum monthly sample data from 1988 through 1994, and the Institut Francais du Petrole monthly sample data from 1991 through 1994. The combination of the large number of sample sets and the fact that each sample fuel is tested by more than 30 laboratories, provides a comprehensive picture of the precision achievable using this test method. Analyzed graphically, the respective sample fuel standard deviations were plotted versus O.N. The variation in precision with respect to O.N. level, for the range of these data, is best expressed by a linear regression of the values. Between 90 and 100 Research O.N., the reproducibility standard deviation is 0.25 unaffected by octane level. This average standard deviation has been multiplied by 2.772 to obtain the limit value.
18.1.1.5 Sample fuels containing oxygenate (alcohols or ethers), in the concentrations typical of commercial spark-ignition engine fuels, have been included in the exchange programs and the precision for these sample fuels is statistically indistinguishable from non-oxygenated fuels in the Research O.N. range from 90.0 to 100.0.
18.1.1.6 The equivalence of this test method when performed at barometric pressures less than 94.6 kPa (28.0 in. of Hg) has not been determined. Reproducibility for the 88.0 to 98.0 Research O.N. range at altitude locations, based on ASTM Rocky Mountain Regional Group interlaboratory test results, would, in the long run, in the normal operation of the test method, exceed approximately 1.0 O.N. only in one case in twenty.
18.1.2 Below 90.0 Research O.N. Range:
18.1.2.1 Precision cannot be stated for the range below 90.0 Research O.N. because current data are not available.
18.1.3 Above 100.0 Research O.N. Range:
18.1.3.1 A limited amount of data above 100 Research O.N. have been obtained by the ASTM Aviation National Exchange Group, Institute of Petroleum, and Institut Francais du Petrole in recent years. Reproducibility for the 101.0 to 108 O.N. range, would, in the long run, in the normal operation of the test method, exceed the values in Table 5 only in one case in twenty.
18.1.3.2 Precision cannot be stated for the range above 108 Research O.N. because current data are not available.
18.2 Bracketing - Dynamic Fuel Level Procedure B:
18.2.1 The amount of data for the bracketing - dynamic fuel level procedure is limited. The available information includes a statistical study involving single ratings by seven laboratories that comparatively tested four gasoline samples and three TSF blends, in the 90.0 to 100.0 research O.N. range, by both the bracketing-dynamic fuel level procedure and the bracketing-equilibrium fuel level procedure. A second phase examined repeatability using duplicate bracketing–dynamic fuel level procedure ratings by each of four laboratories on eight sample fuels.
18.2.1.1 Repeatability of the bracketing–dynamic fuel level procedure is similar to that of the bracketing–equilibrium fuel level procedure as inferred from the statistical analysis of the duplicate ratings data set.
18.2.1.2 Reproducibility of the bracketing–dynamic fuel level procedure is indistinguishable from that of the bracketing-equilibrium fuel level procedure based on the statistical analysis of the limited data from the round-robin study.
18.3 Standard Deviation:
18.3.1 Examination of interlaboratory test results for Research O.N. has been carried out since the late 1930s by the Motor National Exchange Group that regularly tests at least one sample per month. These historical data have demonstrated that the variability (standard deviation) of the test method changes with O.N. as shown in Fig. 8. The curve for this figure is based on ASTM National Exchange Group data from 1966 through 1987.
18.4 Bias - The procedures in this test method for Research O.N. of spark-ignition engine fuel have no bias because the value of Research O.N. can be defined only in terms of this test method.
19. Keywords
19.1 guide table; isooctane; knock intensity; n-heptane; research octane number; spark-ignition engine fuel performance; toluene standardization fuel
