Wind farm performance: why standard monitoring often misses the real picture

Summary

• Wind: the missing reference
• Why existing tools have limits
• Independent, physics-based performance benchmarking
• When this type of analysis is relevant
• Transforming analysis into operational value

Wind: the missing reference

Wind farm performance is harder to track than it appears. In solar energy, asset managers are used to having a Performance Ratio – actual production vs available irradiance. Wind has no equivalent, because the wind received by each turbine cannot be measured independently and reliably. Without an independent wind reference, the split between what the wind did and what the turbine did remains uncertain in most performance report.

The default wind measurement on every turbine (the nacelle anemometer) sits behind the rotor, where airflow is heavily disturbed by the blades. Manufacturers apply correction factors to compensate, known as nacelle transfer functions. But these have several limitations: they are calibrated for turbine control rather than accuracy, established during limited measurement campaigns, and may not hold across different sites, operational modes, or software versions. Critically, they are not traceable once in operation: if a turbine’s output drifts below its reference curve, the correction itself may be adjusted rather than the underlying issue, making the problem invisible rather than solved.

The consequence is that SCADA monitoring cannot reliably tell apart a genuine turbine issue from a wind resource deficit. A turbine that ranks last may simply be receiving less wind than its neighbours. The real underperformers go undetected.

This is where Greensolver’s approach to asset management makes a difference. Rather than developing every capability in-house, Greensolver builds partnerships with best-in-class specialists for each technical domain, and integrates their findings into a single, coherent view of the asset. The partnership with Tipspeed, a company specializing in physics-based wind and performance modeling, is one such example.

Why existing tools have limits

External sensors (masts and lidars) provide accurate, independent wind measurements, but cover one location at a time. Deploying them across a full site is costly and slow. Machine learning methods improve but remain constrained by the same biased nacelle data they are trained on. Atmospheric reanalysis models such as ERA5 offer site-level wind indices, useful to detect large and long-term trends, but too coarse to assess individual turbines reliably.

In short: high-accuracy tools cover too few turbines; broad-coverage tools lack the precision to draw firm conclusions at turbine level.

Independent, physics-based performance benchmarking

Through its partnership with Tipspeed, Greensolver can offer asset owners an independent, physics-based reference for what each turbine should have produced – the wind equivalent of a Performance Ratio. Tipspeed has assessed 2+ GW across 30+ sites. On roughly half, correctable issues were identified (yaw misalignment, controller settings, curtailment errors) and clients have already recovered 31 GWh of annual production. On the other half, the gap was explained by resource or external factors, such new neighboring wind farms. That clarity is equally valuable: it separates wind variance from turbine performance in reporting, and avoids wasted investigations.

At the core of the approach is a digital twin of the wind farm. Using terrain models, satellite imagery, and atmospheric reanalysis data, combined with physical modeling of wind flow, wake deficit, and site-specific performance; the system produces a reference wind speed and expected power output for each turbine, allowing actual production to be benchmarked against a trustable reference.

This analysis can be performed retrospectively, typically over the last years of operation, or implemented as part of an ongoing monitoring solution. By comparing real turbine behavior with the modeled reference, the analysis identifies key performance deviations and loss drivers at both turbine and wind farm level.

Among the most common issues detected are yaw misalignment, controller (pitch or RPM) settings issues, badly implemented curtailments, and wake-related losses. Several documented case studies are available with typical findings – including a site where a lidar was installed to diagnose the turbine ranked worst by SCADA, which turned out to be the farm’s best performer, simply because it received less wind than the others.

Deliverables include turbine-level time series, detailed performance reports with gap analysis to budget and prioritized actions, and access to a visualization platform – giving asset owners the evidence needed to act, or to explain.

When this type of analysis is relevant

Every asset manager should be able to answer two questions: is my farm capturing its full resource? And if not, what is correctable? This is precisely what the digital twin analysis addresses. Ideally, every asset would be monitored continuously, but a one-off analysis is often the right starting point, before moving to ongoing monitoring.

In practice, certain situations make this analysis particularly valuable:

– Suspected underperformance: when production is below budget and the cause has not been identified, or when existing investigations have not led to actionable conclusions.

– Operational changes: when a neighboring wind farm comes online, an upgrade is installed, or when significant operational modifications are made. These events alter the reference conditions of the site and require reassessment.

– Commissioning: detecting issues early – before they compound – and establishing a reliable baseline against which future deviations can be measured, when the OEM is most likely to react fast

– Asset transactions: before or after an acquisition, to establish an independent view of actual performance and surface any issues not visible in standard due diligence.

– Long-term monitoring: ensuring that no drift in performance goes undetected over time, and that corrective actions, once implemented, have delivered the expected gain.

Transforming analysis into operational value

Identifying a performance issue is one step. Getting it investigated, acknowledged, and corrected is another. It requires time, credible evidence, and persistent follow-up. This is where Greensolver’s role becomes operational: coordinating with Tipspeed on the analysis, providing complementary services (site inspections, OEM engagement, ongoing monitoring) needed to move from diagnosis to verified resolution.

Where corrective action is warranted, Greensolver coordinates its implementation, manages discussions with the turbine manufacturer, and verifies that the expected gain has been achieved. Where the analysis reveals that the gap is driven by wind resource or external factors rather than turbine issues, it provides the basis for recalibrating budget expectations and reframing discussions with asset owners and investors.

Greensolver provides the operational and contractual support needed to turn analytical findings into implemented, verified improvements.

Written by Tanya Jacob