Crude Blending — Engineering Fundamentals

Why density blends linearly but RVP, viscosity, and pour point don't — Chevron VPBI, Refutas, PPBI, ASTM D341.

1. Why blending indices?

For an ideal mixture, only properties that are linear functions of mass or volume mix without correction. Density and heat capacity blend linearly on a mass basis. Vapor pressure, viscosity, and pour point are highly nonlinear — a 10% addition of light naphtha to bitumen drops the bitumen viscosity by ~50% and raises its vapor pressure by ~3×.

Industry handles this by transforming each property into a blending index that does mix linearly (on either mass or volume basis), then inverting the index back to the property. Each property uses a different index function tuned to empirical mixing data.

2. Density / API blending

Density blends linearly on a volume basis because density × volume = mass and mass is conserved:

SGblend = Σ(Vi · SGi) / Σ Vi

°API does not blend linearly because the API-to-SG transformation is hyperbolic. Always convert API → SG first, blend, then convert back: °APIblend = 141.5/SGblend − 131.5.

3. RVP — Chevron VPBI

Reid Vapor Pressure is the bubble-point at 100°F for a vented sample. Mixing low- and high-RVP streams gives a result that is below the arithmetic mean. Chevron's VPBI captures this:

VPBI = RVP1.25

Blend volume-weighted, then invert:

VPBIblend = Σ xvol,i · RVPi1.25 ; RVPblend = VPBIblend1/1.25

The 1.25 exponent comes from regressed fits across crude families (ASTM D6378 informational annex). It is conservative — it slightly under-predicts the actual blend RVP, which is what a tariff designer wants.

4. Viscosity — Refutas blending

Refutas (1942) is the industry-standard for crude viscosity blending. Kinematic viscosity (cSt) follows:

VBI = 14.534 · ln(ln(ν + 0.8)) + 10.975

The double-log captures viscosity's exponential temperature dependence. Blend on a mass basis (mass is the conserved quantity for viscous transport):

VBIblend = Σ xmass,i · VBIi

Invert:

νblend = exp(exp((VBIblend − 10.975) / 14.534)) − 0.8

For viscosity at any temperature, ASTM D341 (Walther-MacCoull):

log10(log10(ν + 0.7)) = A − B · log10(TK)

Fit A, B from two known (ν, T) points (typically 100°F and 210°F per ASTM D445). Then extrapolate to pipeline T.

5. Pour-point blending

Pour point (the temperature at which the crude stops flowing) is dominated by wax crystallization. Adding even a small amount of low-pour diluent dramatically depresses the blend pour. The Pour Point Blending Index (PPBI) commonly used in midstream:

PPBI = Tpp(K)12.5

The 12.5 exponent gives very heavy weight to the lowest-pour component, matching field experience that 10% diluent drops the blend pour by 30–50 °F. Volume-weighted:

PPBIblend = Σ xvol,i · PPBIi ; Tpp,blend(K) = PPBIblend1/12.5

6. Dilbit / heavy-crude example

Alberta bitumen (8°API, RVP 2 psi, 10,000 cSt @ 100°F, pour +80°F) blended with natural-gas-plant condensate (60°API, RVP 12, 0.7 cSt, pour −60°F):

70% Bit / 30% Cond50% / 50%Pure Bit
API20.434.68
RVP (psi)~5.3~7.52
Viscosity @ 100°F (cSt)~80~2510,000
Pour point (°F)~+50~+30+80

This is why dilbit shippers target 30% condensate addition: it brings viscosity below the typical 250 cSt pipeline limit and pour point below the line minimum, while keeping RVP under the 10–14 psi summer/winter spec.

7. References

  • ASTM D6378 — Vapor Pressure Measurement by Mini-Method (replaces D323).
  • ASTM D7152 — Standard Practice for Calculating Viscosity of a Petroleum Blend.
  • ASTM D341 — Standard Practice for Viscosity-Temperature Charts.
  • ASTM D445 — Kinematic Viscosity of Transparent and Opaque Liquids.
  • ASTM D97 — Pour Point of Petroleum Products.
  • API MPMS Ch. 12.3 — Calculation of Properties of Blended Petroleum Products.
  • Refutas, J.M. (1942). Viscosity blending index method.
  • Riazi, M.R. (2005). Characterization and Properties of Petroleum Fractions, ASTM MNL50.
  • Wauquier, J.P. (1995). Petroleum Refining Vol. 1: Crude Oil, Petroleum Products, Process Flowsheets, Editions Technip.

Run a blend

→ Open Calculator