ASTM D4056 Standard Test Method for Estimation of Solubility of Water in Hydrocarbon and Aliphatic Ester Lubricants
7. Precision and Bias
7.1 Precision - The repeatability and reproducibility intervals arise entirely from those of the experimental methods employed for n, S, CA and CN. The intervals due to the stated precision of Test Method D1218 will both be 0.4 % of the calculated solubility. The effects of the intervals stated for Test Methods D94 and D3238 are such that no blanket statement can be made. However, the intervals for any specific lubricant can readily be calculated by processing the values of S, CA and CN through the equations alone and plus the intervals from the experimental methods.
7.1.1 Predictions have been made for all pure hydrocarbons for which solubilities of water at 298 K were listed, excepting the olefins, by API, Riddick, and Polak. On aliphatics, Polak's values were chosen if available since they were carefully compared against the older data, giving 14 points. Riddick yielded three, and API two accepted, plus six rejected simply because they appeared "wild". All 18 points were predicted within +/- 10 %. The API and Riddick values were similar on naphthenes; five values were predicted within 20 % and one within 50 %. Polak's six values for aromatics were predicted within 10 %, and five others from Riddick within 15 %. API had only one that was neither redundant nor wild; it was predicted within 10 %.
7.1.2 Predictions at elevated temperatures had to be checked against API data, and were within 10 % of the correct slope from 298 to 373 K for the three liquids covering this range. Most of the data stop at 323 K or less, and have not been analyzed.
7.1.3 API lists eight commercial oils with solubilities at elevated temperatures. Assigning plausible δd, CA and CN values gave predictions on five of these within 50 % or better for solubility, and 25 % or better for slope. The gasoline temperature range was too short to test, and the JP-3 and 4 data appeared wild.
7.1.4 Data by Cantley on a petroleum (SAE 5W) oil, a synthetic hydrocarbon, and two esters at 311 K were predicted within 20 %, and the dependence on RH within 10 %. Other data also illustrate the problems caused by additives.
7.1.5 Riddick shows data on 34 esters. However, only 13 of those meet the criteria in the scope, and two of these were classified as wild because of obvious conflicts with close isomers. The remaining 11 were predicted within 50 % of the true value, and most within 30 %.
7.2 Bias - No general statement is made for bias for this test method since the data used to determine the correlation cannot be compared with accepted reference material.
8. Keywords
8.1 lubricants; solubility; water
