Rooftop Performance and Electricity
Installed Solar Capacity Is Not Daily Energy
Understand why a kW rating does not tell you how many kWh a roof will generate.
You need to size or verify a real system using “Installed Solar Capacity Is Not Daily Energy”, with units, duty and limits that match the application.
Separate a solar array's rated kilowatts from its measured kilowatt-hours, then learn how weather, losses and a fair before-and-after test determine real energy yield.
- Select or check first
- Sunlight at the module plane. Location, season, hour, clouds, tilt, azimuth, horizon and nearby obstructions determine the irradiance reaching the front or rear collecting surfaces.
- Controlling assumption
- Irradiance, weather, module construction, wiring, orientation and measurement boundary match the case being discussed.
- Top risk
- A larger meter total is not yet a causal proof. First confirm the same unit, duration and electrical boundary. Then account for irradiance, temperature, availability and every concurrent system change. Only the remaining, uncertainty-aware difference can reasonably be associated with the upgrade.
- Proof
- Availability and data quality. Shutdowns, faults, maintenance, communications gaps and bad time stamps can reduce real production or make a good system appear to perform badly.
Start with the basics Optional foundation before “Installed Solar Capacity Is Not Daily Energy”
Solar-energy basics
These terms separate what sunlight supplies, what a photovoltaic system converts and what a meter finally records.
- Power and energy
- Power is the instantaneous rate of using or producing energy, measured in watts. Energy accumulates over time and is commonly measured in watt-hours or kilowatt-hours.
- Irradiance
- Irradiance is solar power arriving per unit area, usually expressed in watts per square metre. It is an input condition, not the same thing as electrical output.
- Voltage and current
- Voltage is electrical potential difference; current is the rate of electric charge flow. Their product gives electrical power for a DC operating point.
- Cell, module and string
- A cell is one photovoltaic device, a module connects many cells, and a string connects modules in series. A limitation at one level can affect the others through the wiring.
- Rated versus real output
- Nameplate values come from defined test conditions. Real output changes with sunlight, temperature, angle, shading, wiring, conversion losses and ageing.
How to use this article: Before using a number, identify whether it describes sunlight, DC module output, inverter output or billed energy, and record the time and weather boundary.
On this page
A rooftop described as a 10 kW solar system is not promising 10 kWh every hour, or any fixed number of kilowatt-hours each day. The kilowatt figure is a power rating under defined conditions. Energy is the power actually delivered and accumulated over time. Sunlight, module temperature, orientation, shade, conversion losses, equipment limits and availability all stand between those two numbers.
Power and energy answer different questions
Power is the rate at which electrical energy is being transferred at an instant. Energy adds that changing power over a period. If output were perfectly steady, energy would be power multiplied by time. Solar output is not steady, so an energy meter or monitoring system repeatedly samples power and accumulates the many small intervals.
Energy over an interval
E = P_average × t
Use the average power over the chosen interval, not the array's nameplate rating, unless the array genuinely held that rating for the whole interval. For a varying profile, the meter effectively adds energy from many short intervals.
- E
- electrical energy in kilowatt-hours (kWh)
- P_average
- average electrical power over the interval in kilowatts (kW)
- t
- elapsed time in hours (h)
An average AC output of 4 kW for 2.5 hours contributes 10 kWh. A 10 kW nameplate does not enter that arithmetic unless measured output actually averaged 10 kW.
| Term | What it describes | Question it can answer |
|---|---|---|
| Module or array DC rating, kW or kWp | The sum of rated maximum power under the stated reference or standard test conditions | How large is the rated DC array? |
| Inverter AC rating, kW | The conversion equipment's rated AC power, subject to its voltage, current, temperature and grid limits | What AC power can this inverter deliver within its ratings? |
| Measured power, kW | The DC or AC power observed at one moment and at a named measurement point | What is the system delivering now? |
| Measured energy, kWh | Power accumulated across a defined hour, day, month or year | How much electricity crossed this measurement point during that period? |
| Specific yield, kWh/kWp | Energy divided by rated DC capacity for a clearly stated period and output boundary | How much energy was delivered per unit of installed array rating? |
Diagram description
A central rooftop array is labelled 10 kW installed. A sunny example on the left shows 12,000 kWh per year, while a cloudy example on the right shows 8,000 kWh per year. Both sides say same panels, different yield. The footer says actual generation depends on site, weather, orientation and system losses.
What turns a nameplate rating into real energy
A useful estimate starts with time-resolved solar resource and the array's actual geometry. A performance model then represents module temperature, electrical behaviour and conversion to AC, while applying project-specific losses and operating limits. The final result depends on both the resource that arrived and the system that received it. Even a sound model is an estimate with assumptions and uncertainty, not a production guarantee.
Important influences between rated capacity and measured energy
- Sunlight at the module plane.Location, season, hour, clouds, tilt, azimuth, horizon and nearby obstructions determine the irradiance reaching the front or rear collecting surfaces.
- Cell temperature.A module can run much hotter than the surrounding air. For common crystalline-silicon modules, higher cell temperature generally lowers available voltage and power relative to otherwise similar cooler conditions.
- Shade, soiling and mismatch.Blocked light, dirt and unequal module operating conditions reduce useful array output in ways that depend on string layout, bypass paths and power-electronics architecture.
- Electrical conversion and wiring.DC cabling, connectors, inverter conversion and AC wiring each introduce losses. Their effect must be measured or modeled at the correct boundary rather than hidden inside one unexplained percentage.
- Clipping, curtailment and export limits.Available DC power may exceed an inverter or connection limit, or the system may be commanded below its available maximum. That energy cannot be inferred from the capacity label.
- Availability and data quality.Shutdowns, faults, maintenance, communications gaps and bad time stamps can reduce real production or make a good system appear to perform badly.
Diagram description
The left roof is labelled existing panels and the right roof improved alignment. An arrow between them says same kW. Below, meters show 2,450 kWh and 3,310 kWh over equal illustrative periods, with an arrow for potential extra kWh. A footer explicitly states that the example values depend on site and conditions.
Changing orientation can change the sunlight received without changing the array's rated kW. But the extra energy depends on the original geometry, location, shading, time profile, weather, electrical constraints and any movement energy or downtime. The graphic therefore poses a measurement question; it does not prove that every realignment or tracker will produce the displayed difference.
Normalize carefully when systems have different sizes
Raw kilowatt-hours are useful for billing and total production, but a larger array will usually produce more energy than a smaller one. Specific yield divides energy by rated DC capacity so differently sized systems can be compared on a common scale. It still does not remove differences in weather, orientation, temperature, availability or measurement boundary.
Specific energy yield
Y = E_AC / P_STC
State the period, the AC energy boundary and the capacity convention. Sandia notes that annual yield changes from year to year with weather, so specific yield is a normalized observation rather than a weather correction.
- Y
- specific yield for the period in kWh/kWp
- E_AC
- measured AC energy for the period in kWh
- P_STC
- rated DC array power at standard test conditions in kWp
A 10 kWp array delivering 13,000 kWh of AC energy in one year has a specific yield of 1,300 kWh/kWp for that year and boundary. Another year may differ because the weather and availability differ.
Is performance ratio the same as efficiency?
No. A photovoltaic performance ratio compares final system yield with a reference yield derived from solar irradiation at the array plane and a reference irradiance. It is a system-level monitoring metric, not the cell or module conversion efficiency printed on a datasheet.
Performance ratio helps separate system behaviour from the amount of sunlight, but it still needs consistent sensors, time intervals, data filters and definitions. It is sensitive to temperature unless an agreed correction is applied. A higher or lower value is not automatically proof of one component's quality without reviewing the operating context.
A fair before-and-after test needs more than equal dates
Diagram description
The baseline panel lists 32.5 kWh production, 0.8 kWh auxiliary use or losses and 31.7 kWh net output. The after-upgrade panel lists 43.6 kWh production, 1.1 kWh automation use or losses and 42.5 kWh net output. The stated illustrative difference is 10.8 kWh, or 34.1%, with notices to account for irradiance, temperature and system changes.
Net-energy change in the illustration
ΔE_net = (43.6 − 1.1) − (32.5 − 0.8) = 10.8 kWh
The displayed arithmetic is internally consistent. The percentage is 10.8 divided by 31.7, approximately 34.1%. These are fabricated teaching values; the calculation does not prove a 34.1% gain for any installation.
- ΔE_net
- change in net energy across the two illustrative periods
- production
- gross measured energy at the chosen generation boundary
- auxiliary energy
- energy consumed by relevant controls, drives or other defined auxiliaries
If the before and after periods received different solar irradiation, the 10.8 kWh raw difference mixes the design change with the weather difference. A valid conclusion needs weather-aware analysis.
Build a defensible upgrade comparison
- 1. Freeze the question and boundary.Decide whether the outcome is DC energy, inverter AC energy, energy at the utility meter, self-consumed energy or bill value. Record exactly where every meter sits.
- 2. Establish a clean baseline.Collect interval energy, irradiance, temperatures, equipment states, alarms and downtime before the change. Confirm time zones, clock synchronization, sensor calibration and missing-data rules.
- 3. Record every system change.Document orientation, wiring, inverter settings, cleaning, repairs, firmware, export controls and auxiliary loads. Otherwise several changes can be mistakenly credited to one upgrade.
- 4. Compare equivalent operating conditions.Use an agreed performance model or normalized metric to compare measured energy with expected energy under the actual weather. Equal-duration periods alone are not enough.
- 5. Filter transparently.Identify outages, clipping, curtailment, sensor faults and unavailable periods. State whether each event is included, corrected or excluded and why.
- 6. Report uncertainty and repeatability.Give the observation period, data coverage, model inputs and uncertainty. Prefer enough representative data to show that the apparent change persists beyond one unusual day.
| Weak comparison | Hidden influence | Better evidence |
|---|---|---|
| One sunny day after versus one cloudy day before | Different solar resource | Irradiance-aware expected-versus-measured analysis over representative periods |
| Summer after versus winter before | Sun path, day length and module temperature all changed | A model using actual weather and geometry, or matched seasonal periods across sufficient data |
| Inverter portal before versus utility bill after | Different meters and electrical boundaries | The same calibrated measurement boundary on both sides |
| Gross production before versus net production after | Auxiliary consumption counted on only one side | Subtract defined auxiliary energy consistently from both cases |
| Total kWh from different array sizes | Capacity changed as well as performance | Report total energy, specific yield and all capacity changes together |
| A short test with missing alarms | Downtime, curtailment or data gaps can look like an energy effect | Quality-controlled interval data with event logs and declared filters |
Use the number that matches the decision
Use rated kW to check array and equipment sizing. Use interval kW to diagnose operating behaviour. Use metered kWh to understand production and billing over time. Use specific yield or a weather-aware expected-versus-measured metric to compare performance. For the environmental causes behind lost output, continue with Why Shade and Heat Reduce Solar Output. To compare more modules with improved orientation, read More Panels or Better Panel Orientation?. To connect generated energy to meter units and charges, read From Electricity Units to a Real Solar Bill.
One-minute questions
Questions to answer before you apply this lesson
What should I select or check first for Installed Solar Capacity Is Not Daily Energy?
Sunlight at the module plane. Location, season, hour, clouds, tilt, azimuth, horizon and nearby obstructions determine the irradiance reaching the front or rear collecting surfaces.
Which assumption can change the conclusion?
Irradiance, weather, module construction, wiring, orientation and measurement boundary match the case being discussed.
What can go wrong if I copy the example directly?
A larger meter total is not yet a causal proof. First confirm the same unit, duration and electrical boundary. Then account for irradiance, temperature, availability and every concurrent system change. Only the remaining, uncertainty-aware difference can reasonably be associated with the upgrade.
How do I verify the result in the real system?
Availability and data quality. Shutdowns, faults, maintenance, communications gaps and bad time stamps can reduce real production or make a good system appear to perform badly.
Sources and further reading
- Photovoltaic System Design and Energy YieldU.S. Department of Energy · 2026-10-12Reviewed for the definition of energy yield and the effects of heat, dirt, shade and other system-loss mechanisms on electricity harvested in operation.
- Annual YieldSandia National Laboratories PV Performance Modeling Collaborative · 2026-10-12Reviewed for annual energy yield as kilowatt-hours divided by the PV system's DC standard-test-condition rating and for its year-to-year weather dependence.
- Performance RatioSandia National Laboratories PV Performance Modeling Collaborative · 2026-10-12Reviewed for the relationship among final yield, reference yield, plane-of-array irradiation and the temperature sensitivity of performance ratio.
- Weather-Corrected Performance RatioNational Renewable Energy Laboratory · 2026-10-12Reviewed for why photovoltaic performance reflects both system quality and weather, and why a weather-corrected method is preferable when precise comparisons are required.
- PVWatts CalculatorNational Renewable Energy Laboratory · 2026-10-12Reviewed for the current preliminary energy-estimation inputs, modeled system-loss categories and cautions about assumptions, uncertainty and site-specific characteristics.
- IEC 61724-1:2021 Photovoltaic system performance — Part 1: MonitoringInternational Electrotechnical Commission · 2026-10-12Reviewed for the scope of photovoltaic performance-monitoring terminology, equipment, data acquisition, quality checks and calculated performance metrics.
- IEC TS 61724-2:2025 Photovoltaic system performance — Part 2: Power performance index and capacity evaluation methodInternational Electrotechnical Commission · 2026-10-12Reviewed for the principle of comparing measured output under test conditions with expected output from the system design and for identifying boundaries and data filters.