Tracking the tracker in gigawatt-scale

Our research originated within the EU-funded Horizon Supernova project. While small-scale analyses exist, gigawatt-scale operators typically treat recurring tracker issues as site-specific rather than systemic. At 3E and Statkraft, with access to multi-gigawatt datasets, there are two key questions:

• How reliable are solar trackers in practice?
• Where do they most commonly fail?

We hope this work encourages industry stakeholders to conduct similar analyses and contribute to improved quality standards and grid stability.

Dataset and methodology

We analysed 64 utility-scale PV plants, representing 2.1GWp DC capacity, with datasets spanning six months to five years. Approximately 80% of the sites are located in Europe, primarily in temperate and Mediterranean climates.

To isolate tracker performance from broader plant-level effects, we developed a dedicated KPI: Tracker Availability, focused strictly on mechanical and control uptime.

Actual tracker angles from Supervisory Control and Data Acquisition (SCADA) systems (1–15 minute resolution) were benchmarked against reference angles modelled using PVLIB, based on as-built parameters such as pitch and ground coverage ratio. Trackers deviating by more than five degrees from the reference angle were classified as unavailable.

Only periods with plane-of-array irradiance above 0W/m² were considered. Wind stow positions and planned maintenance events were treated as available time.

We assessed availability under two operational windows:

• All Tracking—the full daily cycle, including backtracking at dawn and dusk
• Core Tracking—peak production hours only

Data quality constraints

Data quality emerged as a major limiting factor. 15 plants were excluded due to severe issues, including stalled signals, scaling and offset errors and mismatches between angle data and generation profiles

Even after exclusions, the dataset comprised hundreds of thousands of tracker records, sufficient to identify robust performance patterns.

Data gaps were frequent and therefore analysed under two assumptions:

• Conservative scenario—communication gaps treated as unavailability
• Best-case scenario—communication gaps treated as available time

Under the conservative scenario:

• All Tracking: 66% median availability (64% average; range 22–96%)
• Core Tracking: 83% median availability (76% average)

The 17 percentage-point gap highlights weaker performance during backtracking and early morning hours, when overnight faults often persist until manual intervention.

Missing data medians reached:

• 11% for All Tracking
• 5% for Core Tracking

Some plants lost up to 70% of tracker records.

Under the best-case assumption (gaps counted as available), median availability increased to:

• 87% (All Tracking)
• 89% (Core Tracking)

These values remain significantly below the 99% availability typically assumed in financial models.

Beyond communication gaps, structural data issues further reduced confidence. Scaling and offset errors distorted angle profiles, while as-built limits sometimes conflicted with observed motion (e.g. ±65 degrees versus documented ±60 degrees). Year-to-year shifts in tracker alignment hinted at calibration drift or wear.

String-level power data helped validate performance on clear days, but tracker angle data proved the most reliable indicator of mechanical behaviour.

Why tracker availability matters

The low levelised cost of electricity (LCOE) of solar PV is strongly linked to tracker deployment. Yet tracker performance is often overlooked in contracts, where performance ratio (PR) and inverter availability dominate.

In an environment of tightening margins, even small availability losses materially affect revenue.

Our Tracker Availability KPI isolates mechanical and control faults from optimisation adjustments. Potential improvements include:

• Prioritising tracker angle and log monitoring
• Automated alerts for deviations exceeding five degrees
• Redundant communication systems
• Machine-learning-enhanced backtracking optimisation

High-density layouts can strain communication networks, while ageing plants face increasing mechanical wear. Embedding tracker availability requirements into contracts, aligned with emerging guidance from certain organisations like the International Energy Agency (IEA), would reduce ambiguity and strengthen accountability.

Toward standardised tracker performance metrics

This work is among the first to systematically evaluate tracker availability across such a broad portfolio of PV plants.

By introducing a simple yet robust KPI methodology, the study aims to enable better communication between manufacturers, engineering, procurement and contracting companies (EPCs) and asset owners and ultimately to enhance operational transparency in the solar industry.

As the sector moves toward tighter margins and higher performance expectations, understanding and quantifying tracker performance becomes a key element in reducing uncertainty, improving energy yield, and ensuring bankable assets.

The views expressed in this article are those of the authors and do not necessarily reflect the views of 3E, Statkraft or its affiliates.