Your partner in production optimization and digital twins.
Your plant drifts. Your model shouldn’t.
Self-learning digital clone that optimize in real time — simple to use, understand and maintain.
Physics Where It Counts · ML Where It Helps◆ AspenTech Independent Solution Provider (ISP) · Presented at AspenTech Optimize 2026
What your twin sees that your instruments can’t
A calibrated first-principles model runs alongside the plant. It matches the measured tags, then derives what no analyzer provides — hydraulic margins, performance against design, fouling — and raises the alarm before a problem reaches operators.
The full digital clone covers every unit in your plant, with real-time optimization, what-if studies and alarm management, running on your own data.
Book a live demoFrom the wellhead to refined products
Follow the hydrocarbon through the plant and see what we plug in at every stage.
What sets our application apart: first-principles highly sophisticated modeling technology and sound physics, self-learning recalibration to live plant data every cycle, and closed-loop targets sent back to the plant continuously.
Not sure where to start? Pick one unit. We scope a pilot with clear acceptance criteria.
Live installations, worldwide
Every live installation on one interactive map. Fly into a region, filter by sector, and open any project to see the work and the value it delivered.
One model. Three jobs. Zero decay.
Built on equation-oriented Aspen HYSYS, the approach that makes plant-wide optimization tractable. The model runs an automated 20–30 minute cycle: reconcile, calibrate, optimize, publish.
- Ingest — historian data (IP.21 / PI) validated and reconciled automatically.
- Calibrate — self-learning reconciliation model configured with an innovative
method that re-tunes parameters to the current plant state.Rolling-Window Modal Standard Deviation (RWMSD)Rolling-Window Modal Standard Deviation (RWMSD) — a short-term variability metric derived from the modal value of a sliding-window standard deviation distribution, giving a stable, outlier-resistant estimate for each tag.
- Optimize — the advanced mathematical model finds the most profitable operating point within constraints.
- Publish — targets and KPIs go to dashboards and APC every cycle.
- What-if studies — test feed, operating or design changes on the calibrated model before they reach the plant.
Model acceptance: Material balance: as good as client plant data · Uptime ≥90% · Temperature, Flow: as per RWMSD
Data & visualize: Data Historian · Dashboard technology of client’s choice
Built to stay right, long after go-live
Most plant models are accurate on delivery day, then drift. Ours keep pace with the plant.
Physics first, not curve-fits
First-principles EO models stay valid outside the data they were tuned on. ML is added to complement.
Accuracy that doesn’t decay
Every cycle the model reconciles to live plant data and re-tunes itself, as catalysts age, feeds change and equipment fouls.
Works with your stack
Built on Aspen HYSYS, Aspen Plus, ACM and Aspen Online, with MySep for separation, and connected to your historian and dashboards.
Proven in live operation
25+ real-time models in 10 countries, averaging about 1,600 instrument tags each, across LNG, refining, gas processing and petrochemicals.
How an engagement runs
A clear path from first data review to a model that keeps paying back.
Assess
Review process data, tags and economics. Agree KPIs and the value case.
→ Scope & value caseBuild
Build a first-principles model of each unit; validate against heat and material balance.
→ Validated modelCalibrate
Reconcile to plant data and tune until the model meets the acceptance criteria.
→ Plant-matched twinDeploy
Run it online, advisory or closed-loop, publishing targets and KPIs to dashboards and APC.
→ Live optimizationSustain
Monitor, re-tune and extend with what-if studies.
→ Lasting resultsIndustries we serve
Select an industry to see what we model — and where we have already done it.
Decades of frontline process expertise
The name Anukoolan means optimization.
“Optimization isn’t a project for us. It’s in our name, our pulse. Your plant moves, the model observes and learns. Then the model leads, and the plant follows. That’s where the money is made.”
Dr. Hiren Shethna
Founder & Optimization Lead- M.Tech IIT Bombay · PhD University of Manchester
- 26 years at AspenTech and Saudi Aramco
- Implemented 25+ digital twin models worldwide
Dr. Karuna Potdar
Simulation & Modelling Lead- B.Chem.Eng + PhD, ICT Mumbai
- 32 years at Reliance Industries
- Led RIL's simulation & modelling team
- 15+ plant models deployed across RIL
Shailesh Sakarkar
Training & Simulation Lead- B.Tech, Laxmi Narayan Institute, Nagpur
- 30 years in the process industry
- 80+ simulation training programs delivered globally
CDU Digital Twin Built with Equation-Oriented Modelling in Aspen HYSYS and Custom Modeling Technology Helps Refinery Improve Margins
Co-authored with AspenTech.
Development of a High-Fidelity FCC Riser Model Through ACM Integration for INDMAX Technology
Co-authored with IOCL R&D Centre Faridabad and AspenTech — a proprietary 27-lump kinetic model validated against commercial plant data.
Technology partners we build on
We deliver on best-in-class process technology, working alongside leading software and equipment specialists.
AspenTech
Independent Solution Provider (ISP)Our models are built on the Aspen suite: Aspen Plus, Aspen HYSYS, Aspen Custom Modeler (ACM) and Aspen Online, delivered as an AspenTech Independent Solution Provider.
Visit aspentech.com ↗MySep
Separation Technology PartnerRigorous gas–liquid separator design and rating, paired with our process simulation for reliable separation-system performance.
Visit mysep.com ↗Knowledge sharing
Your plant is already producing the data.
Tell us about a unit you want to run better. We will scope how a calibrated, self-learning model could optimize it and what that could be worth.
Talk to us
Unit type, the problem, and whether plant data or a model already exists — three lines are enough to start.
Start small
We can scope a single-unit pilot with defined acceptance criteria from day one, to prove value before any wider rollout.