Every reservoir simulation, material balance and surface network calculation rests on one quiet assumption: that the fluid description is right. Whether you use black-oil tables or a full equation of state is a modelling decision, not a software setting, and the best
petroleum software for fluid modelling supports both so you can pick deliberately. This thread explains the difference and gives a simple way to choose.
Quick answer: Use black-oil tables when the fluid behaves simply and pressure stays above bubble or dew point. Use an equation-of-state (compositional) model when phase behaviour matters, such as volatile oil, gas condensate, gas injection or CO2 floods.
What is a black-oil fluid model?
A black-oil model represents the fluid as oil, gas and water, with properties such as formation volume factors, solution gas-oil ratio and viscosity tabulated against pressure. It is efficient and works well for many conventional oils where composition does not change much during depletion.
What is an equation-of-state model?
An equation-of-state (EoS) model describes the fluid as a set of components with defined properties, and predicts phase equilibrium from pressure, temperature and composition. It captures behaviour that tables cannot, such as retrograde condensation or changing composition during gas injection, at the cost of more data and more compute.

Fluid modelling tools such as Kappa Carbone work with both EoS and black-oil representations.
Which one should you use?
Conventional oil above bubble point, waterflood: Black-oil
Volatile oil or gas condensate: EoS / compositional
Miscible gas or CO2 injection: EoS / compositional
Fast screening or large field model: Black-oil (often derived from an EoS)
A workflow that links the two
1. Build and tune an EoS model against laboratory PVT data.
2. Use that model to generate black-oil tables where the simpler approach is acceptable.
3. Compare predictions from both over the expected pressure range.
4. Document where the black-oil approximation breaks down.
What to look for in the software
Clear regression tools to match laboratory experiments.
Consistency checks so tables do not produce non-physical results.
Export formats your simulator can read directly.
Visual diagnostics such as phase envelopes and property plots.
Spotting when your fluid model is the real problem
If a simulation cannot match gas-oil ratio or saturation pressure behaviour no matter how rock properties are adjusted, suspect the fluid model before blaming geology. Check whether the tables extend to the pressures actually reached, whether the fluid sample was representative, and whether a single PVT region is being forced onto a field with composition gradients. Many long history-matching efforts end when the fluid description is corrected. A quick test is to compare simulated and laboratory saturation pressure at the start of the model. If they disagree, no amount of permeability tuning will give a trustworthy forecast.
FAQ
Is EoS always more accurate than black-oil?
Only if it is properly tuned. A poorly matched EoS can be worse than a well-prepared black-oil table.
Can I switch models mid-project?
You can, but it is easier to decide early because it affects data needs, run times and calibration effort.
Over to you: What is your rule of thumb for deciding when black-oil is no longer good enough?