🎓 Lesson 21
D5
Setting Up EOS in Aspen HYSYS: Best Practices & Pitfalls
Setting up an Equation of State (EOS) in Aspen HYSYS means telling the software how fluids behave under different pressures and temperatures so it can accurately simulate real chemical processes.
🎯 Learning Objectives
- ✓ Select and justify an appropriate EOS (e.g., PR, SRK, CPA) for a given feed composition and operating condition
- ✓ Configure pure-component parameters and binary interaction parameters (kij) in Aspen HYSYS using experimental or predictive methods
- ✓ Analyze simulation convergence behavior and diagnose EOS-related errors (e.g., 'No valid phase envelope found')
- ✓ Compare vapor-liquid equilibrium (VLE) predictions from two EOS models against experimental data and quantify %AARD
- ✓ Apply industry-standard kij correlation methods (e.g., Chueh–Prausnitz) to refine mixture behavior in low-data scenarios
📖 Why This Matters
In mining and metallurgical process design—such as pressure oxidation (POX) of refractory gold ores or acid leaching circuits—accurate prediction of phase behavior (e.g., CO₂ solubility in sulfuric acid solutions, O₂/air partitioning in slurry reactors, or H₂S vapor pressure in off-gas streams) directly impacts equipment sizing, safety margins, and environmental compliance. A misconfigured EOS can lead to 20–40% errors in flash calculations, resulting in undersized condensers, overpressurized vessels, or failed relief system designs. This lesson bridges thermodynamic theory with industrial simulation practice—where 'clicking next' without understanding leads to costly field rework.
📘 Core Principles
EOS selection begins with fluid classification: nonpolar hydrocarbons favor cubic EOS (Peng–Robinson, Soave–Redlich–Kwong); polar/water-containing systems (e.g., H₂SO₄–H₂O–O₂–CO₂) require advanced models like CPA or electrolyte NRTL combined with a cubic EOS. Pure-component parameters (critical T/P, acentric factor ω, sometimes volume shift) must be validated against NIST TRC data. For mixtures, binary interaction parameters (kij) correct for non-ideal cross-term interactions; they are *not* universal—they depend on temperature, pressure, and EOS type. Mixing rules (e.g., van der Waals one-fluid) and volume translation (for PR/SRK) further impact liquid density and dew-point accuracy—critical for acid concentration control in leach tanks.
📐 Binary Interaction Parameter Estimation (Chueh–Prausnitz)
When experimental VLE data is unavailable, the Chueh–Prausnitz method estimates kij using critical properties and acentric factors. It’s widely used in mining hydrometallurgy for acid–gas–water systems where lab data is scarce but process safety demands conservative predictions.
Chueh–Prausnitz kij Correlation
k_{ij} = 0.012 + 0.075(\omega_i - \omega_j)^2 + 0.025(T_{r,i} - T_{r,j})^2Empirical correlation to estimate binary interaction parameters when limited experimental VLE data is available.
Variables:
| Symbol | Name | Unit | Description |
|---|---|---|---|
| k_{ij} | Binary interaction parameter | dimensionless | Adjustment term for cross-coefficient in EOS mixing rules |
| \omega_i, \omega_j | Acentric factors | dimensionless | Measure of molecular non-sphericity and polarity |
| T_{r,i}, T_{r,j} | Reduced temperatures | dimensionless | Ratio of system temperature to critical temperature of component i or j |
Typical Ranges:
Nonpolar hydrocarbon pairs (e.g., C₁–C₄): 0.000 – 0.025
Polar–nonpolar (e.g., CO₂–H₂O): 0.010 – 0.050
Strongly associated systems (e.g., H₂O–H₂SO₄): 0.030 – 0.080
💡 Worked Example
Problem: Estimate kij for the CO₂–H₂O pair at 120°C and 20 bar using the Chueh–Prausnitz method. Given: Tc(CO₂) = 304.1 K, Pc(CO₂) = 73.8 bar, ω(CO₂) = 0.228; Tc(H₂O) = 647.1 K, Pc(H₂O) = 220.6 bar, ω(H₂O) = 0.344.
1.
Step 1: Compute reduced temperature Tr,i = T / Tc,i → Tr(CO₂) = 393.15 / 304.1 = 1.293; Tr(H₂O) = 393.15 / 647.1 = 0.608
2.
Step 2: Apply Chueh–Prausnitz equation: kij = 0.012 + 0.075(ωi − ωj)² + 0.025(Tr,i − Tr,j)² → kij = 0.012 + 0.075(0.228−0.344)² + 0.025(1.293−0.608)²
3.
Step 3: Calculate: (−0.116)² = 0.0135 → 0.075×0.0135 = 0.0010; (0.685)² = 0.469 → 0.025×0.469 = 0.0117; sum = 0.012 + 0.0010 + 0.0117 = 0.0247
Answer:
The estimated kij is 0.025, which falls within the typical range of 0.01–0.05 for CO₂–H₂O in acid leach modeling per SME Guideline 2021.
🏗️ Real-World Application
At the Boddington Gold Mine (Western Australia), Aspen HYSYS simulations of the POX autoclave off-gas system initially used default PR kij values for CO₂–H₂SO₄–H₂O. Simulated CO₂ partial pressure was 18% lower than plant DCS measurements, causing underprediction of required scrubber caustic consumption. Engineers re-ran the simulation using kij = 0.032 (regressed from pilot-plant VLE data at 220°C/35 bar) and added volume translation to PR. Result: CO₂ partial pressure error reduced from −18% to +1.3%, enabling accurate sizing of the thermal oxidizer and meeting EPA NOₓ co-emission limits.