Jungle AI - Canopy

The problem isn't lack of data

Most teams already have SCADA, alarms, dashboards, OEM portals, and spreadsheets. The gap is what still gets missed: weak patterns, below-threshold behaviour, and early degradation that does not trigger a clean alarm.

What teams have today

SCADA values and alarms

Static thresholds

Manual investigation

Technical alerts only

What Canopy adds

Abnormal behaviour below alarm thresholds

Changes that only matter under specific operating conditions

Evidence, severity, and operational context in one place

MWh and revenue context behind the anomaly

Canopy learns normal behaviour.
Then flags what changes

Weak patterns, below-threshold behaviour, and early degradation signals often stay hidden until they become obvious enough to trigger an alarm. By then, the issue may already mean component damage, downtime, or avoidable production loss.

Existing monitoring

Your teams already work with SCADA, alarms, dashboards, OEM portals, and spreadsheets. The signals exist, but weak patterns can still stay buried.

Canopy layer

Canopy learns expected asset behaviour, detects abnormal patterns below fixed thresholds and connects them to severity evidence and impact.

Prioritised action

Teams get decision-ready cases with MWh and revenue context, so they know which issues need attention first.

"Thanks to Canopy, we can now proactively identify potential issues ahead of our daily operation management tools, preventing downtime that would have otherwise grown increasingly under the surface."

Jesper van Vliet, Technical Asset Manager, Windunie

See Canopy work on
real fleet data

From abnormal behaviour to impact, evidence, and cases your team can act on.

Detect developing issues

Canopy surfaces developing faults, abnormal behaviour, and performance issues across your fleet before they appear in standard alarms or monthly reports.

Flag context-sensitive anomalies

Canopy flags behaviour that deviates from what is expected under current operating and ambient conditions, not just values that cross fixed limits.

Quantify production impact

Detections are connected to MWh and revenue impact, so teams can understand which issues are worth acting on first.

Investigate to the sensor level

Investigate abnormal behaviour with normality curves, model outputs, and sensor-level evidence down to individual component readings.

Analyse fleet performance

Compare asset behaviour across multiple sensors and operating conditions to find underperformance patterns that standard dashboards often leave buried.

Coordinate action through cases

Detections become structured cases with evidence, severity, impact, and discussion history, so teams can align on the next maintenance or operational action.

“Canopy exemplified its predictive power by identifying abnormal turbine behaviour ahead of traditional monitoring tools.”

Justin van der Sman, Technical Manager RE, Ventolines

“We've been able to efficiently identify curtailment periods and accurately calculate losses.”

Francesco Miniello, Predictive Maintenance Manager Wind & Solar, Repsol Renovables

Business impact that shows up
on your bottom line

Canopy turns early detection into measurable value.

Additional AEP

Average additional AEP seen across the customer base.

15-25%

Margin improvement

Potential margin improvement for many renewable assets.

Canopy detected a rotor bearing issue months before the threshold-based SCADA alarm. The team had time to plan the repair, avoid downtime and save significant production.

“Canopy gave us the lead time to make a strategic decision rather than a reactive one. By aligning this repair with the scheduled gearbox work, we didn’t just save on mobilization costs, we protected our availability for the high-wind season. That single insight secured roughly €190k in potential revenue.”

João Sardo

Head of Asset Management, Galp

What Canopy detects
‍across wind and solar assets

Canopy identifies abnormal behaviour across wind and solar assets before it becomes a production loss, repeated alarm, or unplanned maintenance issue.

Wind

See how Canopy detects abnormal behaviour across onshore and offshore wind portfolios.

●

Rotor and generator bearing overheating

●

Gearbox bearing damage and radiator failure

●

Hydraulic system abnormalities

●

Pitch misconfiguration and control errors

●

Grid curtailment losses

●

Drivetrain and nacelle overheating

●

Turbine underperformance patterns

●

Post-maintenance deviations

Explore wind

Solar

See how Canopy detects abnormal behaviour across utility-scale PV portfolios.

●

Inverter overheating and self-curtailment

●

Snow cover and soiling losses

●

Tracker misalignment

●

String-level underperformance

●

Grid curtailment losses

●

Cooling system abnormalities

Explore solar

Built for the teams accountable
for renewable asset performance

Canopy turns early detection into measurable value.

Asset managers

O&M teams

Performance engineers

Operations teams

Technical directors

Finance and leadership

Asset managers

See which issues threaten availability, production, and revenue across your portfolio.

“Canopy gave us the lead time to make a strategic decision rather than a reactive one. By aligning this repair with the scheduled gearbox work, we protected our availability for the high-wind season.”

João Sardo, Head of Asset Management, Galp

O&M teams

Focus maintenance effort on the behaviours most likely to become downtime or production loss.

“Without Canopy, we would not have identified the issue until much later, after significant losses had already occurred.”

John McGillivray, Director, Africoast

Performance engineers

Investigate abnormal behaviour with model outputs, normality curves, and sensor-level evidence down to individual component readings.

“Canopy exemplified its predictive power by identifying abnormal turbine behaviour ahead of traditional monitoring tools.”

Justin van der Sman, Technical Manager RE, Ventolines

Operations teams

Reduce alarm noise and focus attention on the issues that need action now.

“Thanks to Canopy, we can now proactively identify potential issues ahead of our daily operation management tools, preventing downtime that would have otherwise grown increasingly under the surface.”

Jesper van Vliet, Technical Asset Manager, Windunie

Technical directors

Evaluate asset risk and implementation feasibility with read-only access, no hardware changes, and remote deployment.

“Canopy has proven to be a useful tool for anomaly detection and for the impact quantification of performance-related issues.”

Francesco Miniello, Predictive Maintenance Manager Wind & Solar, Repsol Renovables

Finance and leadership

Connect technical issues to MWh, revenue context, and the business impact of operational decisions.

“Within the TotalEnergies On Program, Jungle has been providing relevant insights in solar and wind farms across 3 different continents.”

Sophie Mouligneau, Program Director, TotalEnergies

Designed to fit into
existing operations

Canopy deploys remotely through a secure, read-only data connection.
It uses the data your assets already generate, requires no new hardware, and does not access live control loops.

Connect existing data

We securely connect to your existing systems and start ingesting data

Train normal behaviour

Our AI learns how your assets behave under all operating conditions.

Go live

Your team gets access to detections, impact views, and cases from day one.

Work with Jungle

We help you act on insights through cases, collaboration, and measurable impact.

No new hardware

•

Read-only access

•

Remote deployment

•

Typically live in 2-3 weeks

Frequently asked questions

Do we need to install specific sensors to get meaningful insights?

No, Canopy doesn't require special datasets or labeled input data. It leverages raw data from your existing sensors, eliminating the need for new sensor installations.

How long does it take to get started?

Typically, Canopy is up and running within two to three weeks, ensuring a swift and efficient deployment process and delivering value from day one.

How much data do you require?

Our models use historical sensor data to learn normal behaviour. In most cases, one year of historical data is enough to make very accurate predictions on the health and performance of your machines.

Do you provide Canopy integrations?

Yes! More and more customers have voiced the request to interact with our normality models and detections via API. Get in touch with our team if this is also of interest to your organisation.

Can you work with any type of machine?

Yes! Canopy's machine agnostic architecture allows it to work with a variety of machines. Our models learn unsupervised, which enables them to harness raw SCADA data, ensuring compatibility with different machine types and operational environments.

Case studies

How Galp turned a potential winter failure into a strategic summer win.

Canopy detected early signs of overheating in the rotor main bearing of a turbine in Valegrande windfarm, avoiding up to €190k in losses.

Find out more

Recovering 126 MWh for SwifterwinT

Detecting post-maintenance pitch errors for the recovery of lost energy.

Find out more

From data to action: Africoast's rapid response to gearbox radiator failure

Canopy enabled a swift response to gearbox radiator failure, restoring turbine performance and avoiding prolonged downtime.

Find out more