Overview

Capacity testing validates that a PV system performs as expected under real-world conditions. Two ASTM standards work together:

  • ASTM E2939 defines expected performance using historical weather data and modeling.
  • ASTM E2848 measures real capacity from monitored data and normalizes to those expectations.

Think of E2939 as “set the yardstick,” and E2848 as “measure against it.”

Why Irradiance and Temperature Aren’t the Whole Story

In a perfect world, power would scale cleanly with irradiance and module temperature. In the field, several factors complicate this:

  • Module temperature isn’t directly measured; ambient temp and wind are proxies.
  • AC output is downstream of multiple losses (wiring, combiners, inverters, transformers).
  • Inverter clipping breaks the “more sun = more power” assumption at high irradiance.
  • Low or unstable irradiance introduces variability and noise.
  • Sparse sensors and coarse intervals (1–5 min) can miss cloud dynamics.

Consequences: you need careful data filtering and a site-specific model. ASTM defines standards for doing this reliably and transparently.

E2939 vs. E2848 at a Glance

E2939 (before testing): Build a regression-based baseline with realistic reporting conditions using long-term weather and simulation.

E2848 (during testing): Collect field data, filter it, then normalize against the E2939 model to determine capacity under reference reporting conditions.

Building the Baseline with PVsyst and 8760 (E2939)

E2939 typically uses PVsyst (or equivalent) with long-term weather files to produce an 8760 (hourly) dataset. From that data:

  • Define reporting conditions by binning & filtering irradiance and temperature in realistic ranges. The goal is to avoid low, unstable, or clipping-prone conditions.
  • Optionally, set reporting conditions for a time period corresponding to the actual test period (e.g., May+June data for a test in June) so summer tests aren’t judged by winter assumptions.
  • Run a multivariable regression of AC power vs. weather variables.
  • Save the regression as the baseline.

This becomes the benchmark against which actual performance is measured during ASTM E2848 testing.

Field Test Preparation & Execution (E2848)

Before collecting data:

  • Calibrate and verify sensors; ensure POA sensors are co-planar and clean. Time-sync all meters, sensors, and inverters.
  • Choose a clear, stable irradiance window; confirm normal plant operation.
  • If possible, do a dry run to validate filters and detect issues.

Performing E2848: Data, Filtering, Normalization

Collect data at the lowest possible time interval (typically 1–5mins). Required data:

  • AC output
  • POA irradiance
  • module and ambient temperature (wind where available)

Filter out:

  • Intervals outside reporting conditions
  • Clipping periods, curtailments, shading events
  • Unstable irradiance (define thresholds appropriate to sensor density and site size).
  • Use professional judgment for filtering thresholds. Larger sites with few sensors need broader instability filters due to partial cloud coverage effects.

Normalize and analyze:

  • Apply the E2939 regression model to normalize the measured data
  • Calculate normalized capacity (i.e. the actual performance the system delivers under the pre-defined reporting conditions)
  • Finally compare this value against the contractual expectations to assess compliance.

Common Pitfalls and How to Avoid Them

Pitfalls:

  • Misaligned or dirty POA sensors
  • Clipping in the test window
  • Unsynchronized timestamps
  • Inclusion of unstable or non-representative intervals

Best practices:

  • Test in stable, non-clipping conditions.
  • Validate with a quick pre-test data review.
  • Use recently calibrated, high-quality sensors.
  • Document assumptions, filters, and anomalies clearly for third-party review.

Reporting and Documentation: What Matters

A clear report should include:

  • Method summary and all filter criteria
  • Time-series plots (irradiance, temperature, AC output)
  • E2939 regression details
  • E2848 normalized capacity results
  • Sensor specs, calibration certificates
  • Explanations for exclusions and any anomalies

Brief Note on Commercial Context: Projects often have predefined performance expectations or reporting conditions. Align your baseline assumptions with those expectations without over-indexing on contract minutiae; the goal is a fair, technically sound comparison based on realistic site conditions.

Conclusion

Accurate capacity testing pairs a realistic, regression-based baseline (E2939) with disciplined data collection, filtering, and normalization (E2848). Focus on sensor quality, rigorous filtering, and transparent reporting to produce credible results that reflect how the plant truly performs.