Thermo-Cognitive Advanced Process Control (TC-APC) models, simulates and optimizes your process in real time to improve yield, conversion and throughput, and reduce energy use, with a quick payback.
Food, dairy and edible oil
Moisture, oil and solvent leave with the product, and the plant finds out at the laboratory.
Dryers, fryers, evaporators and desolventizers run to a setting.
What the product actually carries away is measured after the batch.
Problems we hear
In the words plants use when they search.
- Hexane loss per tonne
- Hexane loss solvent extraction
- Hexane consumption solvent plant
- Milk powder moisture variation
- Milk powder bulk density
- Spray dryer milk powder energy
- Frying oil uptake snacks
- Fryer oil consumption high
- Potato chips oil content
- Rice mill energy saving
- Parboiling steam consumption
- Paddy dryer fuel consumption
Engineering proposal. Begins with a free check.
The science under TC-APC
A dryer's result is decided in the droplet.
A milk powder's moisture, bulk density and particle size are set in the droplet, in seconds. That's the science TC-APC is built on, and AI makes it sharper.
This is the field of our Chief Scientific Advisor, Prof. Pratim Biswas, Member, US National Academy of Engineering and former Dean of Engineering, University of Miami. His published work includes:
- Porous particles made by spray drying. Advanced Materials, 2007
- Predicting the size and charge of particles formed as droplets evaporate. Chemical Engineering Science, 2021
- A model of particle-laden gas flows with strong heat and mass transfer between the gas and the particles. Computers & Fluids, 2006
What nobody measures.
Moisture leaving the dryer
Powder and grain are dried to be safe, which means drier than the specification and heavier on fuel.
Oil in the chip or the cake
Oil uptake and oil left in the cake move with the raw material, not with the setpoint.
Hexane leaving with the meal
Solvent loss per tonne is the number every extraction plant watches and none measures live.
Each of these is worked out live from the physics of the readings your plant already has. TC-APC then holds the process at the optimum as it moves, writing setpoints back to your DCS, PLC or SLC within limits you set.
The nearest proof pages are spray dryers, for milk and food powders held close to their moisture limit instead of well inside it, spice oleoresins and natural extracts, stripped of solvent to specification and no further, protein gelation, surimi to crab stick, where the paste is read from the machine itself, and Plant Ledger.
What we commit to
The same eight commitments, on every project.
No gain, no project
If the check finds nothing worth doing, we say so and stop there.
Savings on your own meters
Before and after, on your instruments, so the result is yours to check.
Shadow mode first
It writes setpoints only after you have watched it be right.
Interlocks untouched
Trips, interlocks and the safety system stay under their own logic.
Runs inside your plant
On its own industrial system and your network, working when the internet is not.
Uses existing instruments
It reads what your DCS, PLC or SLC already measures, and says so where one instrument would close a gap.
Physics first, then AI
A model of the process's own physics, with learning only for what physics cannot know.
Published in Chemical Weekly
The approach, set out as a Special Report in October 2025.
The mark
Find the optimum. Hold it while it moves.
The curve is a yield or selectivity response. The amber point is the optimum, and the dashed drop is where it sits on the operating axis.
The second half of the line is the part that matters. Finding an optimum is a study, and studies end up in drawers. Holding one while the feed, the fouling, the weather and the wear keep moving it is a control problem, and that is the work.
Start with your log sheets
Send one month of log sheets from one area
Photographs are fine. We will show you what the area is losing, which meters disagree with the physics, and what that is worth, including the answer that there is nothing worth pursuing. No charge, and no commitment either way.
Or start with three months of data
If you already keep a historian
Send three months of operating data from one unit, with the laboratory results beside it. The same exercise, done faster, with the same honest answer at the end. How both work.
projects@yielddynamics.in Talk to IIT Bombay engineers Call +91 94230 07004Monday to Saturday, 9:00 to 19:00 IST