Process Control
The MtJ Chemistry Core
Four connected reaction stages, with AI process models tuned to feedstock and reactor geometry, convert methanol into jet-range hydrocarbons at commercial yields.
Reaction Pathway
Four-Stage MtJ Pathway
Methanol-to-jet synthesis follows four interlinked catalytic steps. Metafuels evaluates them separately while representing heat and composition effects between stages.
Methanol Dehydration to Dimethyl Ether
250-350 C / alumina catalyst / WHSV 2-8 h-1
A solid acid turns methanol into DME, producing only water. An AI predictor forecasts catalyst activity loss from inlet temperature plus space velocity, enabling feed-forward bed renewal timing before yield declines.
DME to Olefins via Zeolite Cracking
400-500 C / SAPO-34 / selectivity window 78-85%
Shape-selective SAPO-34 zeolite converts DME into C2-C4 olefins. An adaptive thermal controller preserves the selectivity window as coking reduces activity, lowering the off-spec fraction of C1 by up to 22%.
Oligomerization to Jet-Range Distillate
130-200 C / Ni-SSZ-24 / C8-C16 target fraction
Acid-driven oligomerization joins light olefins into jet-range C8-C16 chains. Feedback control of pressure and temperature raises target distillate yield while limiting heavy wax that otherwise needs downstream hydrocracking.
Hydrotreatment and Isomerization
300-380 C / Pt-alumina / 40-80 bar H2
Hydrotreating at the final stage strips remaining oxygenates and enhances cold-flow behavior. The Metafuels optimizer for hydrogen use trades freeze-point compliance against H2 expense, crucial where green hydrogen availability is limited.
AI Process Loop
Closed-Loop AI Intelligence
Every 90 seconds, AI reads 380+ sensor channels, simulates the next operating hour in a digital twin, and sends new setpoints to DCS controllers. The loop needs no operator.
Static rule controllers presume steady operation; the Metafuels model instead continually recalculates catalyst activity, methanol feed purity, and utility pressure. It offsets departures before they reach product quality.
Feedstock Range
For Biomethanol and Green Methanol
MtJ facilities seldom see uniform feed quality. Biomethanol made through anaerobic digestion has fluctuating water content. E-methanol from power-to-X may contain trace formate tied to the CO2 source. Such changes move reaction equilibrium and alter catalyst life.
Metafuels operates an online feed characterization model that retunes setpoints as incoming methanol quality shifts, drawing on existing tags for conductivity, density, and NIR spectroscopy across most modern plants. Commissioning needs no extra analyzers.
Feedstock origins: biomethanol (agricultural waste, MSW), e-methanol (CO2 hydrogenation), natural gas methanol (for blends during transition), plus blends. At onboarding, the catalyst model is tuned per origin.
Results
Metafuels vs Conventional
Aggregated from internal benchmark data across six MtJ pilot campaigns, each with a minimum 90-day shadow-mode observation period running alongside PID baseline control.
| Parameter | Conventional PID | Metafuels AI | Improvement |
|---|---|---|---|
| Jet-cut yield (% of feed carbon) | 51.2% | 58.8% | +7.6 pp |
| Catalyst cycle length (days) | 34 | 51 | +50% |
| Hydrogen consumption per tonne SAF | 62 kg H2 | 54 kg H2 | -13% |
| Off-spec product rate | 4.8% | 1.3% | -73% |
| Operator alarm rate (alarms/hour) | 14.2 | 3.9 | -73% |
| ASTM D7566 Annex A7 blend compliance rate | 82% | 97.4% | +15.4 pp |
Read the Process Model Brief
Eighteen pages detail model architecture, training-data provenance, and validation against real plant campaigns. Available under NDA to qualified offtake and equipment partners.