A methodology is proposed to predict the thermal conductivity of crude oils (mainly heavy oils) and their fractions based on a distillation assay, asphaltene content, molecular weight, and specific gravity of the fluid. The oils are characterized into a set of pseudo-components and their thermal conductivity is calculated using the Expanded Fluid (EF) thermal conductivity model. The inputs of this model are: the density of the fluid, the pressure, the dilute gas thermal conductivity, and, four parameters that are required for each pseudo-component, rho(s)degrees , lambda(s)degrees, c(2 lambda), and c(3 lambda). The dilute gas thermal conductivity and the parameter rho(s)degrees are calculated from existing correlations. New correlations are proposed for the remaining model parameters and for the binary interaction parameters used in the model mixing rules. The proposed approach was developed and tested on thermal conductivity and density data from the literature for pure hydrocarbons, pure hydrocarbon binaries, bitumen/solvent pseudo-binaries, crude oils, and distillation cuts. In addition, thermal conductivity and density data for pseudo-binaries of C5-asphaltene and toluene were collected in this study at temperatures from 20 to 40 degrees C and pressures up to 10 MPa. The EF thermal conductivity model with correlated fluid-specific parameters predicted the thermal conductivity of 7 crude oils from disparate geographical locations within 3% of the experimental data. Deviations were reduced to within 1% of experimental data by either tuning rho(s)degrees to a viscosity data point or tuning lambda(s)degrees to a thermal conductivity data point.
展开▼
