Many laboratories prefer the PAAR digital density analyzer (ASTM D 4052) above the hydrometer (ASTM D1298) for measuring the relative density of petroleum products. "No significant difference in the true value for the relative density is found on gasoil and jet fuel products. On the contrary for fuel oil the difference between ASTM D1298 and D4052 results has high significance". This is the conclusion from 6 studies performed by IIS on jet fuel, gasoil and fuel oil matrices.

Jet Fuel and Gasoil matrices.

Average values for the relative density are virtually identical. For three different products the values obtained by the ASTM D1298 respectively D4052 are:

Table: Average relative densities at 15oC, measured by two methods on three sample matrices.

The precision parameters found for jet fuel and gasoil comply well with the parameters declared by the respective methods.

Table: Reproducibilities found; (n): number of tests

Jet fuel and gasoil matrices are uncomplicated, with respect to both methods, where D4052 has best precision characteristics.

Fuel Oil

The thick and black fuel oil matrix puts restrictions on the application of the PAAR for density measurement. To overcome the pour point of the oil and to obtain a good fluid, the test temperature is often 50oC, which is beyond the range accepted by D4052. Also the color may hamper the view on air bubbles in the sample. Nevertheless many laboratories appreciate the use of the PAAR also for fuel oils, because of its conveniency and efficiency.

As anticipated, the precision parameters are not as good as for middle distillates, but considerably better than the precision values established for the ‘crude oil’ method ASTM D5002 (R=0.0039 at 0.9500 level).

Table: Average relative densities at 15oC, measured by two methods on three fuel oils.

Reproducibilities were:

Table: Reproducibilities found; (n): number of tests

Interesting is the fact that the Student t-test reveals, that for fuel oil a probability of at least 99% exists that D4052 and D1298 results differ significantly.

Conversion from 50oC to 15oC using table 53B (excluding the glass expansion factor), may not be completely adequate here, maybe due to the reological properties of the oil. Next rounds will be used to study this phenomenon further.