Solar Inverter Performance Indexes

Solar inverter technical performance indexes are the measured behaviors on a datasheet or a certificate, not the adjectives in a brochure. The ones that decide whether a unit may connect are about the grid: does it stop energizing an island, does it trip or ride through, and can it control real and reactive power while doing both. The ones that decide energy yield are about conversion: how much AC comes out across the day’s power range, and how many separate trackers follow strings that are not identical. A single peak efficiency, printed alone, is not either of those answers.

IEEE Std 1547-2018, in the National Renewable Energy Laboratory’s 2022 primer, is the document that fixes the grid half for US interconnection. The unit must detect an unintentional island and cease to energize within 2 seconds, including when grid-support functions are enabled. UL 1741 is described there as the usual US certification path, with type tests run after the support functions so the listing is not “efficient at noon” and “unable to see an outage.” Other countries use other grid codes. An index quoted from a US listing does not travel with the crate.

Grid indexes come before yield indexes

The primer records a priority so the numbers do not contradict each other. Disabling service outranks everything. Tripping on abnormal conditions outranks ride-through. Ride-through may be ended by island detection. Only then come limits on active power and the voltage-regulation functions. A performance sheet that boasts a ride-through curve and is silent on anti-islanding has skipped the index that the standard ranks higher.

The same family of indexes includes how the inverter behaves while it is still connected. The National Laboratory of the Rockies describes the functions that were actually exercised in multi-inverter tests: volt-var with watt priority, frequency-watt, and voltage and frequency ride-through. In those tests, voltage regulation did not harm island detection and sometimes shortened it. Ride-through lengthened detection, and the reported cases still met the 2-second limit. Those results are indexes of a test program, not a rating stamped on every model. The question for a buyer is whether this model’s certificate includes the functions the utility will require, at the settings the utility will lock.

Power factor belongs in the same group. The primer notes that conformance has to be established for different operating modes, including unity power factor, non-unity power factor, and a constant reactive setting, before an islanding judgment is finished. A unit that is only characterized at unity power factor has an incomplete index if it will be asked to inject or absorb reactive power.

Conversion indexes are a curve, not a peak

Conversion efficiency is the AC power out divided by the DC power in, at a stated condition. The condition is the part people drop. A peak at one load, one voltage, and a laboratory temperature is a point. A weighted efficiency, the sort used in listing programs, is an attempt to represent a day in which the inverter spends more time at partial power than at nameplate. This article does not reproduce a weighting formula it has not taken from a primary table. If a seller quotes one percentage, ask which weighting, at which DC voltage, and whether night-time consumption is inside the figure or listed beside it.

Maximum continuous power is a different index from efficiency. It is the AC output the unit can hold, not the highest instantaneous number in a marketing range. Pacific Northwest National Laboratory distinguishes the hardware that does this job: one large central inverter, several central inverters, or many microinverters. Large central units, on that page, use internal transformers. Smaller central inverters and microinverters typically do not. Transformer or transformerless is an architecture index. It changes grounding and it changes whether a module’s potential-induced degradation risk has to be checked, which a 2017 engineering guide on central versus string plants also flagged for floating-ground designs. It is not an efficiency.

How many trackers, and what that does and does not mean

Maximum power point tracking is the control that keeps a DC input near the voltage where that input makes the most power. The count of trackers is an index of independence. The 2017 ICF guide’s comparison table lists central inverters as single or multiple trackers and string inverters as multiple. The point of the extra tracker is practical: strings that differ in shade, tilt, or soiling do not have to share one operating voltage. The guide is careful that a yield gain is site-specific. More trackers do not raise the nameplate efficiency of the power stage. They reduce the loss from forcing unlike strings onto one curve.

A tracker efficiency printed as a percentage is only meaningful with the test behind it. Do not treat a tracking percentage and a conversion percentage as the same kind of number, and do not add them. Ask whether each tracker has its own current limit and its own voltage window, because a tracker that cannot accept the cold open-circuit voltage of the string is not a performance feature. It is a mismatch.

A short list of what to demand on the row

Temperature sits under all of these. A Sandia and NREL study of operating inverters showed component temperatures rising with dissipated power, with the day and the year, and with wind. An efficiency measured at 25 degrees Celsius in a lab is not the efficiency of a cabinet at the temperature the thermal study was written to document. Ask for the power derating curve. How those curves relate to keeping a unit in service longer is the subject of extending inverter life, once that discussion is about heat and settings rather than a slogan.

Harmonic limits and a required power-factor range belong on the certificate for the grid code in force. This article does not copy a percentage from a different standard and call it universal. If the datasheet gives total harmonic distortion, read the power level and the waveform the test used. If it does not, do not fill the blank with a typical figure from another product.

Frequently asked questions

Is the highest efficiency on the brochure the performance index?

It is one point. Yield depends on efficiency across the powers the inverter actually runs, at the DC voltage of the array, including how the unit behaves when it is hot. Ask for the weighted figure and the conditions, not the peak alone.

Does a higher tracker count mean a better inverter?

It means unlike strings can be tracked separately. The 2017 ICF guide treats that as a possible yield gain on irregular sites, not as an automatic gain, and not as a higher conversion efficiency.

Which index decides grid connection?

For US projects following IEEE 1547-2018, the primer puts cease-to-energize, trip, and ride-through ahead of voltage support. The utility’s settings file can be more specific than the headline certificate. A performance claim that ignores that file is incomplete.

Sources and references

Island detection, the 2-second limit, function priority, and power-factor modes in type testing: NREL/Sandia primer on IEEE Std 1547-2018, 2022. Volt-var, frequency-watt, and ride-through test findings: National Laboratory of the Rockies. Central, string, and microinverter architectures, including transformers: Pacific Northwest National Laboratory. Tracker count and the limits of a yield claim: ICF, central versus string inverters, 2017. Temperature versus load and weather: Thermal study of inverter components.