News – Quality Check Your EOS
Why quality-checking your EOS is your secret weapon
02 January 2026
You’ve spent days building your EOS. The data looks clean. The plots look fine. But here’s the truth: many EOS models fail quietly. They seem okay … until you look closer and realize something physically impossible is going on.
Like when your fluid gets heavier but viscosity gets lighter. Or your gas density is higher than your oil density.
That’s not a feature. That’s your model waving a red flag.
Sissel Martinsen nailed this point in her excellent article on how to quality-check a developed EOS. What follows builds on her foundation, and shows how turning EOS validation into a habit gives you an unfair edge.
Like when your fluid gets heavier but viscosity gets lighter. Or your gas density is higher than your oil density.
That’s not a feature. That’s your model waving a red flag.
Sissel Martinsen nailed this point in her excellent article on how to quality-check a developed EOS. What follows builds on her foundation, and shows how turning EOS validation into a habit gives you an unfair edge.
1. Physical consistency: your first line of defense
Martinsen outlined it beautifully: start with the basics.
✅ Liquid density vs molecular weight (or normal boiling point): should increase smoothly.
✅ Viscosity vs molecular weight: same deal, heavier fractions mean thicker fluids.
✅ Boiling point vs carbon number: monotonic, no random dips or spikes.
✅ Critical temperature and pressure vs boiling point: trends that make physical sense.
If you see the opposite, say, density goes up while viscosity goes down, you’re looking at an inconsistent EOS. In reality, heavier fluids resist flow more, not less.
Why it matters: small inconsistencies snowball. An unphysical viscosity trend might not crash your simulator today, but it will slowly corrupt every gas huff-n-puff forecast you make.
2. Thermodynamic consistency: where the real pros stand out
Once your physics check passes, dig deeper.
✅ K-values vs boiling point: single carbon numbers should trend cleanly, no weird crossings.
✅ K-values vs pressure: monotonic trends, crossing lines mean unphysical phase preferences.
✅ Binary interaction parameters (BIPs): check neighbors. Opposite-sign BIPs between similar components? Trouble ahead.
Martinsen emphasized this: if your model predicts oil and gas swapping roles halfway through a pressure range, it’s not predicting, it’s guessing.
3. Build a repeatable quality system
Want to separate yourself from 90 % of engineers? Automate the QA.
That’s how you go from “decent modeler” to “trusted technical lead.”
4. Why positivity and precision matter
People think QA slows you down. It doesn’t. It speeds you up by killing uncertainty.
When you know your EOS behaves correctly:
In Martinsen’s words: “Whether you receive an EOS or you yourself have developed an EOS, it is always a good idea to quality check the EOS.”
Couldn’t agree more.
5. Takeaway
Most EOS models fail not because of bad math—but because no one bothered to check them.
If you want to stop flying blind, use Martinsen’s checklist. Automate it. Baseline it. Bake it into your workflow until quality control becomes second nature.
Do that, and your EOS stops being a mysterious black box, it becomes a trustworthy engine for better decisions.
And if you want your EOS model properly QC’d, reach out to markushays@whitson.com. Markus and the whitson team have turned EOS quality control into an art form.
Martinsen outlined it beautifully: start with the basics.
✅ Liquid density vs molecular weight (or normal boiling point): should increase smoothly.
✅ Viscosity vs molecular weight: same deal, heavier fractions mean thicker fluids.
✅ Boiling point vs carbon number: monotonic, no random dips or spikes.
✅ Critical temperature and pressure vs boiling point: trends that make physical sense.
If you see the opposite, say, density goes up while viscosity goes down, you’re looking at an inconsistent EOS. In reality, heavier fluids resist flow more, not less.
Why it matters: small inconsistencies snowball. An unphysical viscosity trend might not crash your simulator today, but it will slowly corrupt every gas huff-n-puff forecast you make.
2. Thermodynamic consistency: where the real pros stand out
Once your physics check passes, dig deeper.
✅ K-values vs boiling point: single carbon numbers should trend cleanly, no weird crossings.
✅ K-values vs pressure: monotonic trends, crossing lines mean unphysical phase preferences.
✅ Binary interaction parameters (BIPs): check neighbors. Opposite-sign BIPs between similar components? Trouble ahead.
Martinsen emphasized this: if your model predicts oil and gas swapping roles halfway through a pressure range, it’s not predicting, it’s guessing.
3. Build a repeatable quality system
Want to separate yourself from 90 % of engineers? Automate the QA.
- Script it: Automate every plot and check. Humans skip steps, scripts don’t.
- Baseline it: Save your “golden EOS.” Compare new versions against it.
- Explain exceptions: If you keep a non-monotonic trend (e.g., due to real lab data), document it clearly.
- Enforce checks: Make quality validation part of your EOS deployment workflow.
That’s how you go from “decent modeler” to “trusted technical lead.”
4. Why positivity and precision matter
People think QA slows you down. It doesn’t. It speeds you up by killing uncertainty.
When you know your EOS behaves correctly:
- You stop chasing phantom errors.
- You trust your simulation results.
- You save days of debugging.
- You impress clients and management alike.
In Martinsen’s words: “Whether you receive an EOS or you yourself have developed an EOS, it is always a good idea to quality check the EOS.”
Couldn’t agree more.
5. Takeaway
Most EOS models fail not because of bad math—but because no one bothered to check them.
If you want to stop flying blind, use Martinsen’s checklist. Automate it. Baseline it. Bake it into your workflow until quality control becomes second nature.
Do that, and your EOS stops being a mysterious black box, it becomes a trustworthy engine for better decisions.
And if you want your EOS model properly QC’d, reach out to markushays@whitson.com. Markus and the whitson team have turned EOS quality control into an art form.