Action Intelligence™
From insight to action
The system never stops at reporting. Every KPI generates a gap analysis, prioritised actions with named owners, expected outcomes and the measurement that proves it worked.
GE Vernova · Oaxaca Wind Farm, Mexico
Wind Turbine Availability
Oaxaca wind farm at blue hour — the power that keeps the lights on after dinner.
The Gap Analyzer™ · GE Vernova · Oaxaca Wind Farm, Mexico
Wind Turbine Availability
Current state
95%
Target state
98%
Gap
3.0%
Confidence
87%
Analysis chain
Every dashboard answers nine questions
What happened?
Fleet availability held at 95% for the last 14 days, three points below the contracted target of 98%.
Why did it happen?
Downtime is concentrated in nine turbines on the western ridge, where two unplanned gearbox interventions overlapped with a planned inspection window.
What is the gap?
95% today against a target of 98% — a gap of 3.0 points.
How serious is it?
Serious but recoverable. At 95% the site still meets grid commitments, though a further one-point slip would trigger a contractual availability review.
What caused it?
Scheduled maintenance clustering · Unexpected equipment failures · Weather conditions · Supply chain delays · Operational scheduling — presented as hypotheses, tested below.
Who should respond?
Maintenance Team, within 24 Hours.
What actions should be taken today?
Engineering, Operations, Supply Chain have same-day actions.
What happens if we do nothing?
If nothing changes, availability likely drifts toward 93–94% as the inspection backlog grows, roughly 2.1 GWh of clean generation goes unproduced this quarter, and around 6,400 households fall back onto higher-carbon supply.
How will we measure improvement?
Unplanned stop count per 100 turbine-days · Planned downtime hours per week (variance) · Critical spare fill rate …
Possible contributing factors
Hypotheses, not assumptions
Each factor below is supported by a specific piece of operational data. Confidence shows how strongly the data supports it — a human still decides.
Scheduled maintenance clustering
84% supported62% of downtime hours fell inside two planned windows that ran back to back.
Unexpected equipment failures
71% supportedTwo gearbox temperature excursions preceded unplanned stops on turbines 14 and 22.
Weather conditions
58% supportedSustained crosswinds above 19 m/s blocked crane access on four working days.
Supply chain delays
63% supportedTwo pitch-bearing kits are showing an eleven-day lead time against a five-day plan.
Operational scheduling
49% supportedCrew coverage on the western ridge is one technician below the staffing model.
AI decision engine
How the system rates this
- Urgency
- High
- Confidence
- 87%
- Potential business impact
- High
- Potential community impact
- Medium
- Suggested response time
- 24 Hours
- Recommended owner
- Maintenance Team
Confidence reflects how well available operational data supports the leading hypothesis. It is a starting point for a human decision, not a verdict.
Action recommendations
Every recommendation carries an owner, a priority, an outcome, a measurement and a timeline
Engineering
HighReview maintenance records for the nine western-ridge turbines and confirm gearbox wear trend.
- Owner
- Reliability Engineering Lead
- Expected outcome
- Root cause confirmed or ruled out for the two unplanned stops.
- Measurement
- Unplanned stop count per 100 turbine-days
- Timeline
- 24 hours
Operations
HighEvaluate turbine scheduling and resource allocation so planned windows no longer overlap.
- Owner
- Site Operations Manager
- Expected outcome
- Planned downtime hours spread evenly across the month.
- Measurement
- Planned downtime hours per week (variance)
- Timeline
- 3 days
Supply Chain
HighCheck availability of critical replacement parts, especially pitch bearings and gearbox filters.
- Owner
- Regional Parts Planner
- Expected outcome
- Critical spares held on site rather than ordered on failure.
- Measurement
- Critical spare fill rate
- Timeline
- 5 days
Leadership
MediumReview recurring downtime trends across the Latin America fleet in the monthly operating review.
- Owner
- Regional Fleet Director
- Expected outcome
- Systemic pattern identified across sites, not just Oaxaca.
- Measurement
- Fleet availability spread (best vs. worst site)
- Timeline
- 2 weeks
Communications
MediumInform grid and community stakeholders of expected maintenance activity where appropriate.
- Owner
- Community Relations Partner
- Expected outcome
- No surprises for the offtaker or the surrounding municipalities.
- Measurement
- Stakeholder notifications sent ahead of planned work
- Timeline
- 1 week
Learning
MediumShare best practices from higher-performing sites, starting with the 98.6% site in Bahia.
- Owner
- Technical Training Lead
- Expected outcome
- Bahia's pre-window checklist adopted on the western ridge.
- Measurement
- Technicians certified on the new checklist
- Timeline
- 3 weeks
Expected impact
If these actions are successful
- Availability may improve toward the 98% target.
- Reliability increases as gearbox interventions become planned rather than reactive.
- Maintenance efficiency improves through spread windows and on-site spares.
- Operational resilience strengthens across the regional fleet.
- Customer confidence improves ahead of the contract review.
If we do nothing
If nothing changes, availability likely drifts toward 93–94% as the inspection backlog grows, roughly 2.1 GWh of clean generation goes unproduced this quarter, and around 6,400 households fall back onto higher-carbon supply.
Progress tracker
Closing the gap
Current
95%
Gap
3.0%
Target
98%
Same gap, five ways of caring
Who needs to understand this?
Executive view
Wind Turbine Availability
Where the decision is signed — measured in lives, not just line items.
One site is three points short of its availability commitment, with the cause narrowed to clustered maintenance on nine turbines. A 24-hour engineering review and a rescheduled maintenance calendar are the two decisions that matter this week.
AI Improvement Lab™
Every issue becomes something the company learns
Improvement ideas
- Stagger planned maintenance windows so no two overlap on the same ridge.
- Hold pitch bearings and gearbox filters as on-site critical spares.
Innovation suggestions
- Trial a lightweight crane-free access rig for high-crosswind days.
- Pilot acoustic gearbox monitoring alongside temperature sensing.
Automation opportunities
- Auto-alert when gearbox temperature drifts two standard deviations from the fleet median.
- Auto-reorder critical spares when on-site stock falls below a two-week cover.
AI recommendations
- Rebalance the western-ridge crew by one technician for the next 30 days.
- Move the second inspection window nine days later to break the cluster.
Knowledge articles
- Bahia pre-window checklist
- Gearbox thermal drift field guide
Training resources
- Condition monitoring refresher
- Crosswind access safety module
Cross-functional collaboration
- Pair Oaxaca and Bahia site leads for a fortnightly 30-minute exchange.
- Bring Supply Chain into the weekly maintenance planning call.
Success dashboard
How we know it is working
Executive view
Across every open gap today
Top priorities
- Oaxaca wind availability — 3 point gap, 24-hour engineering review
- Manchester imaging uptime — 118 diagnoses rescheduled
- Lisbon overhaul turnaround — four aircraft held into peak season
Highest risks
- A further point of wind availability loss triggers a contractual review
- Imaging backlog compounds into a nine-day waiting-list extension
- Peak-season aircraft availability shortfall for two airline customers
Largest opportunities
- Fleet-wide adoption of the Bahia pre-window checklist
- Predictive coil monitoring across all UK imaging sites
- Digital teardown twin for pre-induction scope accuracy
Projects requiring attention
- Western-ridge crew staffing, one technician below model
- Regional spare coil staging, currently central only
- Test cell utilisation at 97%, no recovery slack
Recognition opportunities
- Oaxaca night crew caught turbine 22 before hard failure
- Manchester radiography absorbed 40 extra scans, zero incidents
- Lisbon test cell crew recovered two days by overnight re-sequencing
Learning opportunities
- Gearbox thermal drift field guide, now fleet-wide
- Artefact recognition module for radiographers
- Borescope accuracy workshop for overhaul technicians
Final rule: every problem ends with a recommendation — and every recommendation names an owner, a priority, an expected outcome, a measurement and a timeline.