PCSAFT Crude Oil Characterization for Modeling of Phase

  • Slides: 27
Download presentation
PC-SAFT Crude Oil Characterization for Modeling of Phase Behavior and Compositional Grading of Asphaltene

PC-SAFT Crude Oil Characterization for Modeling of Phase Behavior and Compositional Grading of Asphaltene Sai R Panuganti, Anju S Kurup, Francisco M Vargas, Walter G Chapman 1

Outline • Asphaltene introduction • Background of asphaltene thermodynamic analysis • Comparison of Cubic

Outline • Asphaltene introduction • Background of asphaltene thermodynamic analysis • Comparison of Cubic and PC-SAFT Eo. S • Robustness of PC-SAFT characterization methodology • Asphaltene compositional grading • Future Work • Conclusion 2

Introduction Asphaltene 1. Polarizable 2. Polydisperse 3. Heavy fraction in crude oil • Operational

Introduction Asphaltene 1. Polarizable 2. Polydisperse 3. Heavy fraction in crude oil • Operational Definition 1. Soluble in aromatic solvents 2. Insoluble in light paraffinic solvents Modified Yen Model 3 Mullins OC. Energy & Fuels 2010; 24(4): 2179 -2207

Modeling Asphaltene Stability Colloidal Model (~1930) Solubility Model (~1980) Stability based on polar-polar interactions.

Modeling Asphaltene Stability Colloidal Model (~1930) Solubility Model (~1980) Stability based on polar-polar interactions. Asphaltene solubilized by the oil. • Micelle formation • Asphaltene particles kept in solution by resins adsorbed on them London dispersion dominate phase behavior. Approaches: (Less parameters) • Flory-Huggins-regular solution theory • Eo. S • Limitations of Colloidal Model: • Negative Hydropilic-Lipophilic Balance for asphaltene [Czarnecki J-2009] • Impedence Analysis – Resins are unlikely to coat asphaltene [Goual-2009] • Diffusion coefficient of asphaltene is same in the presence and absence of resin Nellensteyn FJ. Journal of the Institute of Petroleum Technologist 1928; 14: 134 -138 4

Solubility Model Approaches • Flory-Huggins type models Limitation: 1. Effective molar volume significantly lower

Solubility Model Approaches • Flory-Huggins type models Limitation: 1. Effective molar volume significantly lower than actual molar volume. 2. Cannot account for compressibility • Equations of State 1. Cubic-Eo. S 2. SAFT based models Hirshberg A. Journal of Petroleum Technology 1988; 40(1): 89 -94 5

Modeling using Cubic Eo. S (Crude A) 14000 5% Gas Injection Pressure (Psia) 12000

Modeling using Cubic Eo. S (Crude A) 14000 5% Gas Injection Pressure (Psia) 12000 Exp AOP 10000 Exp Bu p AOP (SRK(P)) 8000 Bu P (SRK(P)) 6000 Crude A 4000 2000 0 0 100 200 300 400 Temperature (F) • Characterized the crude oil system using PVT-Sim of Calsep • The Cubic Eo. S employed was SRK-P 6

Modeling using Cubic Eo. S (Crude A) 14000 5% Gas Injection 12000 Pressure (Psia)

Modeling using Cubic Eo. S (Crude A) 14000 5% Gas Injection 12000 Pressure (Psia) 14000 10000 8000 6000 4000 2000 0 0 100 200 Temperature (F) 30% Gas Injection 12000 300 400 0 0 50 100 150 200 250 300 Temperature (F) The optimized Cubic Eo. S parameters from 5% were used to predict the phase behavior for 30% injected gas Limitations of cubic equation of state: • Asphaltene critical properties are not well known • Results are very sensitive to parameters 7 Larry GC et al. Advances in Thermodynamics (Volume 1): C 7+ Fraction Characterization. Taylor & Francis; 1989

Introduction to SAFT Parameters represent the physical system directly • PC-SAFT EOS is be

Introduction to SAFT Parameters represent the physical system directly • PC-SAFT EOS is be used • Parameters for most compounds are known Chapman WG et al. Industrial Engineering and Chemistry Research 1990; 29(8): 1709 -1721. Gonzalez DL et al. Energy & Fuels 2005; 19(4): 1230 -1234. 8

PC-SAFT Characterization Developed a standardized characterization procedure based on: SARA analysis Flashed Liquid and

PC-SAFT Characterization Developed a standardized characterization procedure based on: SARA analysis Flashed Liquid and Gas compositions (C 9+) Molecular Weights Liquid Density Bubble Pressure AOP Stable Unstable VLE Methodology: • Composition data up to C 9+ is sufficient. • Few parameters were needed • Temperature independent binary interaction parameters for all compounds are very small Panuganti SR et al. “PC-SAFT Characterization of Crude Oils and Modeling of Asphaltene Phase Behavior” Fuel - Submitted 9

Comparison of PC-SAFT and Cubic Eo. S 5% injected gas (A) Pressure (Psia) 12000

Comparison of PC-SAFT and Cubic Eo. S 5% injected gas (A) Pressure (Psia) 12000 SRK-P 9000 PC-SAFT (Crude A) 6000 3000 0 0 200 Temperature (F) 400 Characterized using PC-SAFT and SRK-P Eo. S • Will PC-SAFT work better than Cubic EOS? • Will a specific set of PC-SAFT parameters be sufficient to capture the phase behavior of the system at a different condition? 10

Comparison of PC-SAFT and Cubic Eo. S 5% injected gas (A) Pressure (Psia) 12000

Comparison of PC-SAFT and Cubic Eo. S 5% injected gas (A) Pressure (Psia) 12000 No gas injection (B) SRK-P 9000 PC-SAFT 6000 3000 0 0 200 Temperature (F) 400 0 100 200 300 Temperature (F) 400 Crude B Better performance of PC-SAFT is visible • Will PC-SAFT with proposed characterization procedure be able to predict phase behavior for higher amounts of gas injected? 11

Comparison of PC-SAFT and Cubic Eo. S 5% injected gas (A) Pressure (Psia) 12000

Comparison of PC-SAFT and Cubic Eo. S 5% injected gas (A) Pressure (Psia) 12000 No gas injection (B) SRK-P 9000 PC-SAFT 6000 3000 0 200 400 200 300 Temperature (F) 400 Crude B 15% injected gas (C) 12000 Pressure (Psia) 100 9000 PC-SAFT holds upper hand over C Eo. S 6000 What about for even higher gas injection? 3000 0 0 200 400 Temperature (F) 12

PC-SAFT vs. Optimized Cubic EOS 5% injected gas (A) Pressure (Psia) 12000 No gas

PC-SAFT vs. Optimized Cubic EOS 5% injected gas (A) Pressure (Psia) 12000 No gas injection (B) SRK-P 9000 PC-SAFT 6000 3000 0 0 200 400 0 100 200 300 400 30% injected gas (D) 15% injected gas (C) Crude B Pressure (Psia) 12000 9000 6000 3000 0 0 200 Temperature (F) 400 0 100 200 300 400 Temperture (F) 13

Prediction of Effect of Gas Injection 15% injected gas (A) 15000 Pressure (Psia) 12000

Prediction of Effect of Gas Injection 15% injected gas (A) 15000 Pressure (Psia) 12000 Crude C 9000 6000 3000 0 0 200 400 • A different crude, exhibiting different physical properties. • Characterized using standardized methodology 14

Robust Methodology 15% injected gas (A) 15000 No gas injection (B) Pressure (Psia) 12000

Robust Methodology 15% injected gas (A) 15000 No gas injection (B) Pressure (Psia) 12000 9000 6000 3000 0 0 200 400 0 100 200 300 400 30% injected gas (D) 10% injected gas (C) Pressure (Psia) 12000 Crude C 1. Robust Methodology 2. Good parameter estimation 9000 6000 3000 0 0 200 Temperature (F) 400 0 100 200 300 400 Temperature (F) Any property of the precipitate phase can be calculated 15

Compositional Grading Introduction Used for: 1. To predict oil properties with depth 2. Find

Compositional Grading Introduction Used for: 1. To predict oil properties with depth 2. Find out gas-oil contact compositional How is asphaltene grading useful? Reservoir connectivity A M Schulte. SPE Conference; September 21 -25, 1980 Høier L, Whitson CH. SPE 74714; 2001; 4(6)525 -535 16

Compositional Grading Algorithm Whitson C H & Belery P; SPE 28000 1994 443 -459

Compositional Grading Algorithm Whitson C H & Belery P; SPE 28000 1994 443 -459

Reservoir Compartmentalization Optical Density (@ 1000 nm) 0 0. 5 1 1. 5 2

Reservoir Compartmentalization Optical Density (@ 1000 nm) 0 0. 5 1 1. 5 2 2. 5 3 22500 23000 PC-SAFT (M 21 B) 23500 Depth (ft) 24000 PC-SAFT (M 21 A) 24500 25000 PC-SAFT (M 21 A North) 25500 26000 26500 27000 27500 • All zones belong to the same reservoir as the gradient slopes are nearly the same. • The curves do not overlap meaning each of them belong to different zone. 18

Approximate Analytical Solution ρ= Molar density; h=Depth; Mi =Mol wt = Partial Molar Volume

Approximate Analytical Solution ρ= Molar density; h=Depth; Mi =Mol wt = Partial Molar Volume Assumptions: 1. Changes in density of oil with depth can be neglected 2. At infinite dilution partial molar volume is independent of composition 3. System is far away from critical point such that partial molar volume is independent of pressure changes Sage BH, Lacey WN. Los Angeles Meeting, AIME; October 1938 Morris Muskat. Physical Review 1930; 35(1): 1384 -1392 19

Approximate Analytical Solution Optical Density (@ 1000 nm) 0 0. 5 1 1. 5

Approximate Analytical Solution Optical Density (@ 1000 nm) 0 0. 5 1 1. 5 2 2. 5 3 22500 PC-SAFT (M 21 B) 23000 23500 Depth (ft) 24000 Model (M 21 B) PC-SAFT (M 21 A) 24500 Model (M 21 A) 25000 PC-SAFT (M 21 A North) 25500 Model (M 21 A North) 26000 26500 27000 27500 • We have Partial molar volume of asphaltene = 1934 cm 3/mol. It corresponds to a particle size of 1. 83 nm • Analytical solution can be used for sensitivity analysis and approximate estimate. 20

Future Work • Tar mat occurrence due to compositional grading of asphaltene. • QCM-D

Future Work • Tar mat occurrence due to compositional grading of asphaltene. • QCM-D experiments for determination of asphaltene deposition rates and aging effects. • Micro fluidic studies to understand the asphaltene deposition mechanism. 21

Conclusion • Solubility model using PC-SAFT Eo. S • PC-SAFT characterization methodology proposed •

Conclusion • Solubility model using PC-SAFT Eo. S • PC-SAFT characterization methodology proposed • Robustness of PC-SAFT characterization methodology • Evaluate reservoir compartmentalization through asphaltene compositional grading. 22

Acknowledgement • • Walter G Chapman Francisco Vargas Anju S Kurup Jeff Creek 23

Acknowledgement • • Walter G Chapman Francisco Vargas Anju S Kurup Jeff Creek 23

24

24

Characterization of T Oil 18000 16000 14000 Pressure (Psia) 12000 C 1 Bu P

Characterization of T Oil 18000 16000 14000 Pressure (Psia) 12000 C 1 Bu P 10000 C 1 AOP C 2 Bu P 8000 C 2 AOP 6000 C 3 Bu P C 3 AOP 4000 2000 0 0 10 20 30 40 50 60 70 Amount of precipitating agent added (Mole %) 80 90 100 25

Derivation of Thermodynamic Model 26

Derivation of Thermodynamic Model 26

Algorithm 27

Algorithm 27