Evaluation the Performance of Empirical Correlations for Predicting the Dead Oil Viscosity and Isothermal oil compressibility for Libyan Crude Samples

Abstract

Understanding the PVT parameters is essential for various petroleum calculations, including the assessment of hydrocarbon flow characteristics, forecasting future performance, designing production facilities, and strategizing enhanced oil recovery methods. Ideally, the PVT properties of hydrocarbons are derived from laboratory analyses conducted on either bottom-hole samples or recombined surface samples. However, laboratory data may not always be accessible due to economic or technical constraints. In such instances, empirical correlations are employed to estimate these properties. These correlations have been formulated based on fluid samples from specific geographical regions. Given the diverse compositions of crude oils across different areas, relying on empirical correlations to predict PVT properties may yield unsatisfactory results when applied to hydrocarbons that behave differently from the fluid samples used to develop the correlations. This study assesses various PVT correlations to estimate the viscosity of dead oil (μod) for Libyan crude oils, utilizing 58 data points, and examines isothermal compressibility (Co) for Libyan crudes with 145 data points across approximately 100 distinct reservoirs. The evaluation criteria employed in this research include statistical and graphical error analysis. Existing correlations were applied to the Libyan dataset, and error analysis was conducted by comparing the predicted values with the original experimental values. The most accurate correlation for each PVT parameter has been identified. The findings indicate that the correlation proposed by Kartoatmodjo & Schmidt yields a dead oil viscosity with an Average Absolute Error Ratio (AAER) of 18.53% and a relative coefficient of R² = 0.1512. For isothermal oil compressibility, the correlation by Ahmed (1985) demonstrates the best performance, with an AAER of 31.67% and a relative coefficient of R² = 0.104%.