What-If Scenarios & Optimisation
Fork live state · run perturbations · let Bayesian Optimiser propose recipes · compare on Pareto frontier
Pre-canned scenariosclick to load
Manual perturbation
Station
Cycle Δ -10%
Yield Δ +0.5 pp
Parallel Δ 0
Horizon 15m
Causal sensitivity tornado
Click Run to compute Δ-throughput per station.
Pareto frontier · cost vs throughput24 BO candidates
Bayesian optimiser
Recipe tuning · objective: max η, min scrap
Best so far · iter 13
η 23.18% · scrap 1.21% · firing peak 825°C, dwell 4.2s
Saved scenarios · side-by-side3 variants
| Scenario | Throughput /h | $/W | gCO₂/W | Δ vs base | Action |
|---|---|---|---|---|---|
| Baseline (today) | 1080 | $0.108 | 13.8 | 0 / $0.000 | baseline |
| SMED upgrade | 1142 | $0.103 | 13.6 | +62 / $-0.005 | |
| TOPCon switch | 994 | $0.117 | 14.4 | -86 / +$0.009 |