Harnessing Energy-Efficient Habits with Data in Shipping
Energy efficiency in shipping is ultimately a set of habits — and data is what makes good habits visible, measurable, and repeatable. A virtual conference talk on connecting fleet data to the daily behaviour that determines fuel consumption.
A talk from the Digital Ship online Vessel Performance Forum (remote, during the COVID period), on using operational data to build and sustain energy-efficient habits across a fleet.
What the talk covers
- Fouling is expensive and silent: even slime films can raise resistance and power by ~20%, and heavy calcareous fouling by over 85% — and it builds up while vessels sit idle.
- The "one or two extra tonnes a day" trap: large vessels burn up to ~100 t/day, so small daily tolerances quietly add up to enormous industry-wide waste.
- Data quality first: robust filtering and validation, driven by human experience — high-frequency data is ideal, but well-handled noon reports can still produce reliable results.
- ISO 19030 vs machine learning: the transparent classic methodology alongside neural networks / XGBoost — but models must obey the laws of physics and predict what companies actually need, not just score well on standard error metrics.
- Proof from real vessels: ISO 19030 correctly captured hull-cleaning effects with documented savings up to ~$10,000/day; an ML fouling model trained on one cleaning predicted the next one closely.
- Culture and crew: continuous validation, a clear financial incentive, and crew buy-in (passage-plan updates, generator-start thresholds) — delivered with less workload, more context, and proper training.