Solar Plant Performance & Field Testing

How owners, developers, and independent engineers verify that a solar plant delivers what was promised — from the 7-day ASTM E2848/E2939 capacity test to the one-year IEC 61724-1 performance ratio test, and every sensor and data point that makes those results defensible.

The short version

A performance test is how you convert a construction milestone into a financial fact. Before final payment is released, a warranty clock starts, or a plant changes hands, the owner and the independent engineer need proof that the array actually produces the power the contract promised. For utility-scale PV, that proof rests on two complementary tests: a short, intensive capacity test (ASTM E2848 and E2939, typically run over about seven clear days) that answers “does the plant hit its rated power right now, under measured conditions?”, and a one-year performance ratio (PR) test (IEC 61724-1) that answers “does the plant sustain its efficiency across a full cycle of seasons, temperatures, and soiling?” Both stand or fall on instrumentation: calibrated pyranometers, an albedometer, a soiling station, temperature and wind sensors, a rain gauge, and revenue-grade energy metering. This guide explains each standard, each test, and each instrument and data point in depth for utility-scale photovoltaic plants, including bifacial designs.New Paragraph

Why Field Performance Testing Is the Moment That Matters

A solar project spends years in development and months in construction, but its entire financial premise comes down to one question: will it generate the energy the model predicted? Performance testing at mechanical completion is where that question gets answered with measured data rather than assumptions. The results drive substantial completion and final payment to the EPC, trigger production guarantees and liquidated damages, establish the baseline for long-term operations, and satisfy the lenders' independent engineer (IE) that the asset is bankable.

Getting testing right protects every party. For the owner, it confirms the revenue basis of the investment and prevents a chronically underperforming plant from being accepted. For the EPC, a rigorous, standards-based test provides a fair and defensible way to demonstrate that the work meets contract — and to be paid promptly when it does. For the IE and lenders, it de-risks the debt. When testing is done poorly — with uncalibrated sensors, weak data filtering, or the wrong standard applied — disputes follow, and they are expensive.

This guide walks through a complete, modern testing scope exactly as it appears in real solicitations — a seven-day capacity test and a one-year PR test, backed by a full weather and metering sensor suite. Each standard, each test, and every instrument and data point is explained in depth below.

A typical utility-scale testing scope

Capacity test: ASTM E2848 + ASTM E2939, 7 consecutive days. Performance ratio test: IEC 61724-1, one year. For a bifacial plant, the scope adds rear-POA pyranometers and an albedometer, and in hot, dust-prone climates the back-of-module, soiling, and rainfall measurements become especially consequential — as explained throughout.

The Two Pillars: Capacity Test vs. Performance Ratio Test

Owners frequently conflate these two tests, but they answer different questions and are governed by different standards. The capacity test is a snapshot — a short, statistically rigorous measurement that normalizes production to a fixed set of “reporting conditions” and asks whether the plant delivers its expected power at that instant in its life. The PR test is a movie — a long-duration measurement of how efficiently the plant converts available sunlight into delivered energy across an entire year of real weather.

A well-structured project uses both. The capacity test clears the path to final payment quickly after construction; the PR test confirms, over the following year, that the plant's efficiency holds up in service. Crucially, both tests share the same instrumentation backbone — which is why the two instrument lists in the scope are identical, and why investing in a single, properly designed on-site weather and metering system serves both purposes.

The Capacity Test: ASTM E2848 and ASTM E2939

The capacity test is the industry-standard way to verify a PV plant's power output at commissioning. It is built from two complementary ASTM documents that are almost always cited together: E2848 defines the measurement and regression method, while E2939 defines the reporting conditions and the expected capacity that the measurement is compared against.

ASTM E2848 the test method

ASTM E2848, Standard Test Method for Reporting Photovoltaic Non-Concentrator System Performance (current edition E2848-13(2018)), establishes how to measure a system's AC power output and relate it to the environmental conditions that drive it. Rather than trying to catch the plant at one perfect instant, E2848 collects many short-interval measurements over several days and fits a multivariate regression model that describes power as a function of the weather.

The regression takes the general form:

P = E × ( a₁ + a₂E + a₃Tₐ + a₄v )

where P is the measured AC power, E is the plane-of-array (POA) irradiance, Tₐ is ambient air temperature, v is wind speed, and a₁–a₄ are coefficients found by least-squares fitting to the measured data. The dominant term is irradiance; temperature and wind are second-order corrections that capture how the array warms and cools. Once the coefficients are known, the model can predict the plant's power at any specified set of conditions — which is exactly what makes a fair comparison possible.

Data filtering and quality control



The credibility of the regression depends on excluding data that would distort it. Common, standards-aligned filtering criteria applied during an E2848 test include:

• Irradiance floor: data points below roughly 400 W/m² are excluded, because low-light behavior is non-linear and noisy.
• Stability windows: intervals with rapidly changing irradiance (passing clouds) are removed so that measured power and measured irradiance correspond to the same moment.
• Reporting-condition band: for the final capacity determination, points are typically drawn from within about ±20% of the reporting irradiance so the result reflects representative operating conditions.
• Minimum sample size: a sufficient number of valid intervals (for example, on the order of 50 fifteen-minute points) is required for the regression to be statistically sound.
• Availability accounting: periods when equipment is offline for reasons unrelated to the plant's capability (e.g., grid curtailment) are handled explicitly so the plant is judged on what it can do, not on external constraints.

This is why the scope specifies seven consecutive days: it is usually enough time to accumulate the required volume of clean, high-irradiance, stable data — but if weather is poor, the window may need to extend until the data quality thresholds are met. A capacity test is finished when the data are good enough, not merely when the calendar runs out.

ASTM E2939 reporting conditions and expected capacity

A measured number means nothing without something to compare it to. ASTM E2939, Standard Practice for Determining Reporting Conditions and Expected Capacity for Photovoltaic Non-Concentrator Systems (current edition E2939-13(2018)), supplies that reference. It defines the Reporting Conditions (RC) — a single representative set of irradiance, ambient temperature, and wind speed derived from the site's typical operating profile — and the Expected Capacity, which is the power the plant is predicted to produce at those exact conditions based on the design model.

The reporting conditions are not arbitrary: they are chosen to represent the conditions under which the plant most often operates while producing meaningful power, so that the capacity result is representative rather than cherry-picked from a single ideal moment. The test then compares the two numbers directly:

Capacity Ratio = Measured Capacity (at RC) ÷ Expected Capacity (at RC)

Both the measured and the expected values are evaluated at the same reporting conditions using each side's regression/model, which removes weather as a variable and isolates the plant's true capability. Acceptance is defined by contract, but a capacity ratio of 95% or greater — with regression errors held to within a few percent — is a widely used threshold. Fall short and liquidated damages or corrective work may be triggered; meet it and the plant clears a major commercial hurdle.

Why E2848 and E2939 are always cited together

E2848 tells you how to measure and model the plant; E2939 tells you what conditions to evaluate it at and what to expect. E2848 without E2939 gives you a model with nothing to judge it against; E2939 without E2848 gives you a target with no rigorous measurement. Together they form the complete, defensible capacity test that lenders and EPC contracts rely on.

Instruments required for the capacity test

The E2848 regression is driven by measured weather, so the capacity test requires the full sensor suite listed in the scope — GHI and front-POA pyranometers, rear-POA pyranometers, diffuse pyranometers, an albedometer, a soiling station, an ambient air temperature sensor, wind speed sensors, back-of-module sensors, and a rain gauge — alongside revenue-grade power measurement. Each of these is explained in depth in the Instrumentation section below.

The Performance Ratio Test: IEC 61724-1

Where the capacity test is a commissioning snapshot, the performance ratio (PR) test measures how the plant behaves over a full year of operation. It is governed by IEC 61724-1:2021, Photovoltaic system performance – Part 1: Monitoring, the international standard for how PV monitoring systems are designed, what they measure, and how accurately.

What performance ratio actually measures

Performance ratio expresses how much of the energy that could have been produced (given the sunlight that actually fell on the array and the plant's nameplate rating) was in fact delivered. It is dimensionless, usually expressed as a percentage, and is defined as the ratio of the final yield to the reference yield:

PR = Yf ÷ Yr

The final yield (Yf) is the net energy delivered divided by the plant's rated DC power. The reference yield (Yr)) is the in-plane (POA) irradiation divided by the reference irradiance of 1,000 W/m² at standard test conditions. In plain terms: reference yield is the number of “peak sun hours” the array received; final yield is the number of full-power hours it actually produced. Their ratio is the fraction of ideal performance achieved, after real-world losses such as temperature, soiling, wiring, inverter conversion, mismatch, and downtime.

Temperature-corrected PR

A plant's PR naturally dips in summer because hot cells are less efficient — a physical effect, not a defect. To judge the equipment fairly across seasons, IEC 61724-1 also defines a temperature-corrected performance ratio, which normalizes each interval's output to a reference cell temperature using the modules' temperature coefficient and the measured back-of-module temperature. This is precisely why the scope includes back-of-module sensors: without them, a summer PR guarantee in a hot climate would be unfairly penalized by heat that has nothing to do with build quality. Temperature-corrected PR separates “the modules are hot today” from “the plant is underperforming.”

Monitoring classes and why utility-scale means Class A

IEC 61724-1 defines monitoring classes — A, B, and C — that scale accuracy and sensor requirements to the stakes of the project. Class A is the highest-accuracy tier and the expectation for utility-scale, financed assets. Class A calls for the most accurate, calibrated instruments (for example, ISO 9060 Class A pyranometers), tighter measurement-uncertainty limits, and a fuller set of measured parameters. The 2021 edition notably added explicit treatment of reflected/albedo irradiance and bifacial considerations, and it requires dew and frost mitigation (heating/ventilation) on Class A irradiance sensors so that morning moisture does not corrupt data.

Sampling, recording, and calibration

Class A monitoring specifies how often data is sampled and recorded. Fast-changing quantities such as irradiance and electrical power are sampled at short intervals (on the order of seconds) and typically stored as one-minute averages, which are then aggregated to hourly, daily, monthly, and annual figures for PR reporting. All sensors must carry current, traceable calibration certificates, and the standard sets accuracy and placement requirements for each measured parameter. Rigorous data completeness — accounting for and flagging any gaps — is part of a defensible one-year result.

Why the PR test runs a full year

A one-year duration is not padding; it is the point. Only a full annual cycle captures the summer heat that suppresses efficiency, the winter cold that lifts it, the seasonal sun-angle changes that shift bifacial gain, the wet and dry periods that drive soiling and natural cleaning, and the pollen or dust events characteristic of the site. A month-long test could flatter or punish a plant depending on when it happened to run; a year gives owners and lenders an honest, defensible measure of sustained performance.

Data to be collected — primary parameters

The scope's primary parameters are the measurements that directly enter the PR calculation and the plant's energy accounting. Each is explained below, followed by a reference table.

• i. Global Horizontal Irradiance — GHI (W/m²). The total solar energy striking a horizontal surface. GHI is the reference against which other irradiance measurements are checked, supports satellite/model correlation, and underpins quality control of the POA data.
• ii. Diffuse Horizontal Irradiance (W/m²). The scattered (non-direct) portion of sunlight, measured with a shaded pyranometer. It lets the site's light be decomposed into direct and diffuse components, which is essential for transposition modeling and for understanding performance under variable skies.
• iii. Front Plane-of-Array Irradiance — POA (W/m²). The irradiance in the plane of the modules on the front side. This is the single most important environmental measurement: it is the reference yield's numerator and the dominant driver of production. Multiple sensors are distributed across a large array to represent it faithfully.
• iv. Rear Plane-of-Array Irradiance (W/m²). For bifacial modules, the irradiance reaching the back of the panels. Measuring it directly is how bifacial gain is quantified and credited, rather than merely assumed — central to fairly evaluating this plant.
• v. Ambient Air Temperature (°C). Air temperature in a shielded/aspirated housing. It characterizes site conditions, feeds the E2848 regression, and provides context for module-temperature behavior.
• vi. Albedo (%). The fraction of sunlight reflected by the ground, measured with an albedometer. Albedo largely determines how much light reaches the modules' rear side, so it is a first-order input to bifacial performance.
• vii. Soiling Losses (%). The energy lost to dust, dirt, and debris on the module glass, measured with a soiling station. It distinguishes true equipment underperformance from a surface that simply needs cleaning, and informs the O&M cleaning schedule.
• viii. Wind Speed (m/s). Wind cools the modules and is a correction term in the E2848 regression; it also provides operating context for irradiance sensors.
• ix. Electrical Energy Produced — Wh (energy meter). The revenue-grade measurement of delivered energy. This is the numerator of the plant's real output and the basis of the final yield in PR. Meter accuracy and traceability are as important here as sensor calibration.
  

Data to be collected secondary parameters

Secondary parameters are not always in the headline PR formula, but they are indispensable for diagnosis, temperature correction, and explaining why a plant performs as it does.

• i. Back-of-Module Temperature (°C). The module's actual operating temperature, sensed on the rear surface. It is the basis of temperature-corrected PR and the key to judging the plant fairly through a hot summer.
• ii. DC Power at the Inverter Level (kW). Power on the DC side of each inverter. Comparing DC power to expected values localizes underperformance to specific combiner/inverter blocks and reveals string or tracker issues.
• iii. AC Power at the Inverter Level (kW). Power on the AC side of each inverter. Paired with DC power, it exposes inverter conversion efficiency and clipping behavior.
• iv. Rainfall (mm). Measured with a rain gauge. Rainfall correlates with natural cleaning and soiling recovery, helping interpret soiling trends and validate the cleaning strategy.
  

Instrumentation Deep-Dive: Every Sensor in the Scope

Both the capacity test and the PR test draw on the same instrument suite. Below, each instrument from the scope is explained in depth — what it measures, why it matters, and what good practice requires for accuracy and placement. For a utility-scale, financed, bifacial plant, all of these should meet IEC 61724-1 Class A expectations with current, traceable calibration.

GHI and Front-POA pyranometers

Pyranometers are thermopile sensors that measure broadband solar irradiance. Two orientations are used. A GHI pyranometer is mounted horizontally to capture global horizontal irradiance; front-POA pyranometers are mounted coplanar with the modules to capture the irradiance actually reaching the array's front face. POA is the workhorse measurement — it is the independent variable in the E2848 regression and the reference-yield input in PR — so multiple POA sensors are distributed across a large site to represent spatial variation, and on trackers they move with the modules. For Class A monitoring these should be ISO 9060 Class A (secondary-standard) pyranometers with heating/ventilation to prevent dew and frost, and traceable calibration.

Rear-POA pyranometers

Bifacial modules generate power from light striking their back surface, so rear-POA pyranometers are mounted facing the ground behind the modules to measure that rear irradiance directly. Because rear irradiance varies with row position, module height, ground cover, and shading, several rear sensors are typically deployed and averaged. These measurements are how bifacial gain is credited from data rather than estimated — essential to fairly evaluating a bifacial plant and a capability formalized in the 2021 edition of IEC 61724-1.

Diffuse pyranometers

A diffuse pyranometer measures the scattered component of sunlight by blocking the direct beam — traditionally with a shadow ball on a solar tracker, or with a shadowband. Subtracting diffuse from global (and combining with direct-normal information) lets the site's irradiance be fully characterized, which improves transposition modeling from horizontal to plane-of-array and strengthens quality control of the POA data. It supplies the scope's Diffuse Horizontal Irradiance parameter.

Albedometer

An albedometer is effectively two pyranometers back-to-back: one facing up to measure incoming irradiance and one facing down to measure reflected irradiance. Their ratio is the ground albedo. For bifacial plants this is a first-order measurement, because the rear-side resource depends directly on how reflective the ground is — bright soil, gravel, or vegetation each yield very different bifacial gains. Consistent albedo measurement is what turns rear-side production from an assumption into a verified quantity.

Soiling station

A soiling station quantifies the energy lost to dust and dirt on the modules. The common approach compares two matched reference devices — one kept clean and one allowed to soil naturally — and derives the soiling ratio (and thus soiling losses) from the difference in their output. This matters enormously in semi-arid, dust-prone environments, and it protects both parties: it prevents a dirty array from being mistaken for a defective one, and it gives the owner data to optimize cleaning frequency against cost.

Ambient air temperature sensor

Ambient temperature is measured with a precision sensor (typically a platinum RTD) inside a radiation shield, often aspirated, so that sunlight does not bias the reading. It is a correction variable in the E2848 regression, a characterization parameter for the site, and important context for interpreting module temperatures. Proper shielding and placement away from heat sources are essential to accuracy.

Wind speed sensors

Anemometers measure wind speed at a defined, documented height. Wind convectively cools the modules — raising efficiency — and appears as a term in the E2848 regression, so capturing it improves the model's fidelity. Wind data also provides operating context for the irradiance sensors and the plant's thermal behavior. Both cup and ultrasonic anemometers are used; placement and height must be recorded for the data to be meaningful.

Back-of-module sensors

Back-of-module (BOM) temperature sensors — RTDs or thermocouples bonded to the rear of representative modules — measure the cells' true operating temperature. Because silicon efficiency falls as temperature rises, BOM temperature is the key input for temperature-corrected PR and for normalizing the capacity result. Multiple sensors are placed across the array (different rows, orientations, and positions) to represent the fleet, and good thermal bonding is critical to avoid reading air rather than module temperature.

Rain gauge

A rain gauge — usually a tipping-bucket type — records precipitation depth. Rainfall is the primary natural cleaning mechanism for a PV array, so correlating rain events with the soiling station's readings explains soiling recovery, validates the cleaning strategy, and adds context to seasonal performance trends. It provides the scope's Rainfall parameter.

Bifacial and Site-Specific Considerations

Bifacial performance is only as good as its measurement

This plant's inclusion of rear-POA pyranometers and an albedometer signals a bifacial design, and bifacial plants raise the stakes on instrumentation. Bifacial gain — the extra energy from rear-side light — can be a meaningful share of production, but it is highly sensitive to ground albedo, row spacing, module height, and tracker behavior. If rear irradiance and albedo are not measured well, the plant's true capability is either under-credited (unfair to the EPC) or over-credited (a risk to the owner and lender). Rigorous rear-side and albedo measurement, aligned with IEC 61724-1:2021, is what makes a bifacial guarantee defensible.

Heat, soiling, and grid curtailment

Local climate and grid conditions shape the test in three ways. First, heat: high summer temperatures depress raw PR, making back-of-module measurement and temperature correction essential to a fair result, and a summer test window means the capacity test may run in demanding thermal conditions. Second, soiling: semi-arid or agricultural environments make the soiling station and rain-gauge correlation particularly valuable over the year. Third, grid curtailment: in constrained or volatile markets, curtailment and negative-price intervals can force the plant offline for reasons unrelated to its capability, so availability and curtailment must be carefully separated from performance in both tests to avoid unfairly penalizing the asset.

How Keentel Engineering Delivers Performance Testing

Keentel Engineering supports owners, developers, and independent engineers across the full performance-testing lifecycle — turning a scope like this one into a defensible, on-schedule result. Our involvement typically spans:

• Test plan and protocol development. A project-specific procedure that binds the contract's guarantees to ASTM E2848/E2939 and IEC 61724-1, defines reporting conditions and expected capacity, and sets acceptance criteria and data-filtering rules up front — before disputes can arise.
• Instrumentation design and commissioning. Selecting, siting, and verifying the full sensor suite (POA/GHI/rear/diffuse pyranometers, albedometer, soiling station, temperature, wind, rain, and revenue-grade metering) to Class A standards with traceable calibration.
• Capacity test execution and analysis. Running the seven-day (or longer, as data quality requires) test, performing the regression, applying quality filters, and reporting the measured-to-expected capacity ratio.
• One-year PR monitoring and reporting. Data acquisition, gap accounting, temperature correction, soiling and bifacial analysis, and periodic PR reporting through to the final annual result.
• Independent review and dispute support. Objective, standards-based analysis that owners, EPCs, and lenders can all trust — including review of a counterparty's test where an independent check is needed.



Get a quotation for your performance test

Planning a capacity test or a one-year PR test for a utility-scale or bifacial PV plant? Keentel Engineering can scope the instrumentation, write the test protocol, and execute and analyze both the ASTM E2848/E2939 capacity test and the IEC 61724-1 PR test. Contact us at keentelengineering.com to request a quotation and align the test with your EPC contract and lender requirements.

Frequently Asked Questions