Offshore O&M
Fleet and maintenance scheduling that treats crew rest and safety requirements as constraints inside the optimisation, not as a check performed afterwards.
Start a conversation
The problem
Offshore maintenance is scheduling under weather, vessel availability and crew limits at the same time — and every hour of unnecessary downtime is paid for in lost generation.
Safety requirements such as mandatory rest after a task are usually applied after a plan is built, as a filter. The result is a plan that looks efficient and then falls apart on contact with the rules.
The honest question is what those requirements actually cost in fleet utilisation and turbine availability — and that question is rarely answered with numbers.
What the optimiser does
- Schedules maintenance tasks across a vessel fleet over a planning horizon
- Treats crew-rest requirements as hard constraints inside the model, not as post-hoc filtering
- Uses real weather series rather than idealised conditions
- Minimises makespan while reporting fleet utilisation and turbine availability alongside it
What it lets you measure
- The operational cost of a specific safety requirement, quantified by comparing identical scenarios with and without it
- Where fleet capacity is genuinely the binding constraint and where it is the weather window
- How a change in vessel mix or task mix moves availability
Why it can be checked
- Built on constraint programming with an open solver, not a proprietary black box
- Runs on public wind farm data and published reanalysis weather series
- Delivered as a reproducibility package: data, code, results and figures together
- A third party can rerun it and get the same answer — which is the point
How we work
Constraints belong inside the model
A rule applied after optimisation produces a plan that was never optimal under the rule. Modelling the requirement directly gives you both a feasible plan and its true cost.
Reproducible by default
Results that cannot be independently rerun are opinions. We publish enough for someone else to reproduce the work, and we build on public data where public data exists.
Comparison, not assertion
Two scenarios, identical in workload, fleet, weather and objective, differing only in the constraint under study. That is what makes the number about the constraint and not about our setup.
Questions we are usually asked
Do we have to hand over operational data to see anything?
No. The method is demonstrated on public data and published weather series before your data is involved at all. Only then does it make sense to talk about your fleet.
Which safety requirements can be modelled?
Any requirement that can be stated as a constraint on scheduling — rest periods, qualification requirements, transfer limits, working-time rules. We start from the ones that bind hardest in your operation.
Does this replace our planning system?
It answers questions your planning system is not built to answer — what a rule costs, where capacity binds, what a different fleet would do. It can inform your planning rather than replace it.
Bring us a scheduling question
The most useful start is a concrete one: a fleet, a season, a requirement you suspect is expensive. We will show what the model says.
Start a conversation