2018 年 63 巻 10 号 p. 706-714
A linear equation of viscosity-pressure-temperature-density of lubricants was derived in the former report (1st part), and a linear equation of density-pressure-temperature was derived in 2nd part. If the gradients δ [K/GPa], a [GPa・K] and intercept b [‒] of the respective linear equations are found, it is possible to easily calculate the high pressure viscosity, the high pressure density and the pressure viscosity coefficient. In this report, author investigated whether these three values can be predicted by multiple regression analysis with the molecular structure of lubricant independent of pressure as an explanatory variable. As a result, attention was focused on the molecular weight and the functional group of the molecular structure showing the properties of the lubricant. Each atom as the minimum unit of the segment, that is, primary to quaternary carbon, aromatic carbon, carbonyl carbon, carbonyl oxygen and ether oxygen atoms, or hydrogen atoms bonded to primary to tertiary carbon were used as explanatory variables. Further, each density and viscosity at 40℃ and 100℃, which can be measured at atmospheric pressure, and refractive index,viscosity index, the temperatures at density of 0.75g/cm3 and 0.95g/cm3, furthermore, temperature at viscosity of 107mm2/s were used as explanatory variables. Author derived each multiple regression equation using these explanatory variables.
