Prediction of physical and chemical qualities of crude-oil using statistical time series analysis.

Abstract

An application of a statistical time series analysis method is given for the interpretation of physical and chemical qualities of crude-oil samples taken from two oil-fields in Poland. The analytical data was gathered in the years 1995-1999, the information being collected once a month. Randomly selected crude-oil samples were examined in several analytical laboratories. For each sample examined, the following chemical-physical parameters were defined: date, source of the sample, specific gravity, density, colour, relative viscosity, viscosity, kinematic viscosity, drip point, setting point, etc. The data obtained played a decisive role in classifying the samples of crude-oil and in forecasting the properties of crude-oil that will be obtained from the mines in the future. To process the data obtained, the STATISTICA programme was used. It contains a module for time series analysis and for graphical presentation of the results.