At a glance: Commercial solar monitoring is set up to catch inverter shutdowns and lost communication. The losses that cost owners the most are partial: an open string, a derating inverter, a soiling film that builds in Florida humidity. This article explains how each one appears in the data, compares the metrics that can expose it (performance ratio, performance index, and inverter-to-inverter comparison), lists what changes when the fix happens on an occupied building, and ends with a checklist that assigns each review to a named role.
Commercial solar monitoring protects a Florida array only when someone compares what each inverter produced against what the sunlight and temperature on that day allowed, because the costliest production losses on commercial systems do not trip an alarm. An open string fuse, a failed connector, or a biological soiling film can take output for months while the portal still reports every inverter online. This article explains how to read monitoring data to separate equipment underperformance from weather, which metric to trust for which question, and who should act on the result. It does not rank monitoring platforms or hardware brands, and it does not estimate the dollar loss on any specific building.
The readers who need this most manage arrays that have run for years without a formal production review. Their question is whether the array still delivers what its financial case assumed. The answer usually sits in data the system already collects.
Why does commercial solar monitoring miss so much solar production loss?
Commercial solar monitoring misses production loss when alerts go unanswered and when equipment keeps running at reduced output, because most commissioning setups report only shutdowns and lost communication. kWh Analytics published contributor findings on both problems in its 2026 Solar Risk Assessment, announced May 12, 2026. SmartHelio found that inverter shutdowns, the failure monitoring does flag, drive 28 percent of recoverable solar performance risk across 6.5 GW of assets. A case study from Univers and Lightsource bp traced a 15 percent power loss to undersized fuses, a partial loss that raises no shutdown alarm. Those figures come largely from utility-scale portfolios, so they describe the mechanisms rather than a predicted loss for any one rooftop.
The same report carries a Solar Unsoiled finding that fungal soiling causes a 5 percent loss in the humid subtropical region of the United States, which covers most of Florida. Biological films adhere to module glass more firmly than loose dust, so their signature in the data is a slow decline across many inverters rather than a sudden step, and an alarm-only setup never flags it.
Separating a fault from aging needs a reference rate. National Renewable Energy Laboratory researchers assembled nearly 2,000 measured degradation rates in Photovoltaic Degradation Rates: An Analytical Review and reported a median of 0.5 percent per year. An array producing several percent less than the prior year under similar sunlight has a problem that aging does not explain. The U.S. Department of Energy’s Solar Energy Technologies Office O&M resources tie regular operations and maintenance to meeting production targets, and the monitoring review is where that maintenance gets its work orders.
How does a commercial solar array lose production without tripping an alarm?
A commercial solar array loses production silently when a fault reduces output without stopping the inverter, because most inverter alarms are written for shutdowns and grid events. A string inverter with one open input keeps converting power from its remaining strings, so its output falls by that string’s share and nothing else changes. Across a site with many inverters, that loss looks like a hazy day in the site total until the affected inverter is compared against its neighbors under the same sky.
Heat creates a second silent loss. Inverters reduce output to protect their electronics when internal temperature climbs, and an enclosure in direct afternoon sun on a Florida roof runs hotter than a shaded one. The result is a flattened afternoon curve on one unit, often with a temperature event in a log nobody reads. Clipping looks similar and is a design condition. An array with a high DC-to-AC ratio holds flat at its AC limit on clear middays by intent, and the review should record those hours so they are not mistaken for faults.
| Loss mechanism | What the monitoring data shows | How it is confirmed on site |
|---|---|---|
| Open string fuse or failed connector | One inverter or input reads lower than its neighbors on every clear day, by a near-constant share | Current measurement at the combiner or inverter input; IV curve trace |
| Inverter thermal derating | One unit flattens below its AC rating on hot afternoons, with temperature events in its log | Inverter event log review; enclosure shading and ventilation check |
| Inverter clipping (design condition) | Flat-topped midday curve at the AC rating on clear days across units with the same DC-to-AC ratio | Comparison to the design production model; no repair needed |
| Soiling or biological film | Gradual decline across many inverters, with partial recovery after heavy rain | Visual inspection; production comparison on a test section before and after cleaning |
| New shading | Loss at the same time each day that shifts with the season | Site walk at the affected hours; drone survey |
| Communication loss | Data gaps or flat zeros while the utility meter still records export | Gateway, data logger, and network check |
To spot an underperforming commercial solar array, compare each inverter’s daily energy per kilowatt of installed DC capacity against its neighbors and against the production expected for the irradiance measured on site. A gap that persists in clear weather points to equipment. A gap that appears only in cloudy or hot periods points to the model or to design conditions.
New shading and lost communication complete the list. Tree growth along a carport row or a new rooftop unit set beside an array produces a loss that recurs at the same hour each day. A gateway that loses its network connection leaves the array producing while the data shows nothing, and any fault that starts during the outage waits for the next site visit.
Which monitoring metric shows real solar array underperformance?
No single metric shows every kind of solar array underperformance, so the right setup layers alarms, inverter-to-inverter comparison, and a weather-adjusted index. Sandia National Laboratories’ PV Performance Modeling Collaborative defines the AC performance ratio according to IEC 61724 as the system’s AC yield divided by the reference yield calculated from plane-of-array insolation, and notes that the ratio is typically lower in warm seasons than in cold ones. On a Florida building, a summer dip in performance ratio can be normal temperature behavior rather than a fault.
The performance index described by the same Sandia collaborative compares measured energy to the energy a performance model predicts for the measured weather. For a healthy system it stays close to 1 and does not swing with the seasons, which makes it the better month-over-month health number. Its weakness is the model itself. A model built with the wrong module data, tilt, or shading inputs will hide real losses or invent ones that are not there.
| Method | What it catches | Where it falls short |
|---|---|---|
| Alarm-only monitoring | Inverter shutdowns, grid trips, lost communication | Misses partial losses; an inverter running at reduced output raises no alarm |
| Performance ratio (PR) | Whole-system output against measured plane-of-array irradiance, as defined in IEC 61724 | Moves with temperature and season; needs a working on-site irradiance sensor |
| Performance index (PI) | Measured energy against modeled energy for actual weather; near 1 when healthy | Only as accurate as the model and its weather inputs |
| Inverter-to-inverter comparison (kWh per kWdc) | One inverter or input lagging its neighbors under the same sky | Misses losses that hit every inverter equally, such as uniform soiling |
| String- or module-level data | Locates a fault to a string or module before the truck rolls | Adds hardware, data volume, and communication points that need upkeep |
Inverter-to-inverter comparison is the quickest test because it needs no weather model. Inverters with the same capacity, orientation, and module type should produce nearly the same energy per kilowatt each day, so the outlier is the lead. Shared losses such as soiling still need the index. Owners who want this review done for them can see how the AGT commercial solar operations and maintenance program handles monitoring and reporting.
What drives the size of a solar production loss and how fast it is found?
The size of a solar production loss is the lost output per day multiplied by the days between the fault and the repair, so detection speed matters as much as the fault itself. The drivers below set both terms. Neither this article nor Advanced Green Technologies publishes a dollar value per lost kilowatt-hour, because that value depends on the building’s tariff, its demand charges, and how much of the production is used on site.
| Driver | Why it changes the loss | Source or basis |
|---|---|---|
| Data granularity | Inverter- or string-level data shows a single-string loss that site totals hide | Sandia PVPMC performance metrics |
| On-site irradiance sensor | Performance ratio requires plane-of-array insolation; without it, weather and faults blur together | Sandia PVPMC performance ratio definition |
| Alert routing and dispatch terms | Loss accrues from fault to repair; the O&M contract sets that clock | AGT solar O&M contract guidance |
| Climate | Fungal soiling causes about a 5% loss in the U.S. humid subtropical region | Solar Unsoiled, kWh Analytics 2026 Solar Risk Assessment |
| Module aging | Median degradation of 0.5% per year sets the baseline for year-over-year comparison | Jordan and Kurtz, NREL |
Dispatch terms deserve the closest reading because they are negotiable. An alert that reaches an unwatched inbox adds the same lost days as no alert. The guidance in what a solar O&M contract should cover sets out how alarm thresholds, named recipients, and time to dispatch belong in writing before the first fault.
What changes when a monitoring alert leads to work on an occupied Florida building?
On an occupied building, a monitoring alert starts a coordination process before it starts a repair. The first step is remote confirmation from the data, so a technician arrives knowing which inverter, input, or roof zone is affected. The property manager then receives notice of roof access, the work area, and any inverter shutdown window, because isolating a string or an inverter interrupts that portion of the array’s production for the duration of the work.
Electrical diagnosis on a live array is hazardous work. The OSHA solar energy electrical hazards page notes that solar workers face electrocution and arc flash hazards. Modules produce voltage whenever light reaches them, so string testing is planned around lockout procedures and protective equipment, with the work zone kept clear of occupants.
Roof condition belongs in the same visit. A technician can note ponding or membrane wear under the array, and a roof and drone solar survey can document a large roof without foot traffic across every row. AGT completed a 2.2 MWdc rooftop array at the Orange County Convention Center in 2026 while the building stayed in operation, as Solar Power World reported in April 2026, and service calls on busy buildings need the same scheduling discipline, as described for commercial solar maintenance in Miami.
Which Florida rules and standards apply to commercial solar monitoring as of November 2026?
Commercial solar monitoring data and utility billing data measure different points, and Florida’s net metering rule addresses the second. Under Florida Administrative Code Rule 25-6.065, investor-owned utilities interconnect and net meter customer-owned renewable generation, and net metering lets that generation offset the customer’s on-site consumption. The utility’s meter records net flow at the service point, while the monitoring portal records what the inverters produced, so the two numbers will not match. The utility and the Florida Public Service Commission set how exports are credited, and how FPL credits commercial solar exports covers that process in detail.
IEC 61724-1, the international standard for photovoltaic performance monitoring, defines classes of monitoring systems and addresses sensor selection, installation, and accuracy; its current edition dates from 2021. Replacing an inverter or adding a weather station mount can be permitted work. The Florida Building Code 9th Edition (2026) takes effect December 31, 2026, and the permit application date generally determines which edition governs, so the building official confirms the code path for any equipment change.
Commercial solar monitoring checklist: who confirms what?
A monitoring program works when each review has an owner and a frequency. Each task below can be assigned by name in the O&M contract or the facility’s procedures.
- Facility manager: confirm alerts reach a named person and a backup, and test the path each quarter.
- Facility manager: export twelve months of inverter-level production and the original design model.
- O&M provider: confirm the on-site plane-of-array irradiance sensor is clean, level, and reporting before relying on performance ratio.
- O&M provider: compare each inverter’s monthly kWh per kWdc against its neighbors and report any gap that persists in clear weather.
- O&M provider: record clipping hours and inverter temperature events so design behavior is separated from faults.
- Asset manager: reconcile portal production with utility meter data each quarter and send billing questions to the utility.
- Property manager: schedule roof access and inverter shutdown windows with tenants before field work begins.
- Owner: keep monitoring records and repair logs on file for lender, insurer, and warranty requests.
Underwriters ask about maintenance history on coastal arrays, and what underwriters check on Florida commercial solar explains which documents they review. Coverage questions go to the owner’s insurer or broker.
Frequently Asked Questions
What is commercial solar monitoring?
Commercial solar monitoring is the hardware and software that records production, equipment status, and site conditions on a commercial array and reports them through a portal and alerts. A basic setup reads only inverter output and fault codes. A fuller setup adds a revenue-grade meter, a plane-of-array irradiance sensor, and string-level current, so production can be compared with what the weather allowed.
How do I know if my commercial solar array is underperforming?
Compare each inverter’s daily energy per kilowatt of DC capacity against inverters with the same orientation and module type. One that lags by a steady share on clear days likely has a string, fuse, or connector fault. For whole-array losses, compare measured production with the design model’s expectation for the measured weather. A sustained gap beyond normal aging warrants a site visit.
What is a good performance ratio for a commercial solar system in Florida?
There is no single target, because performance ratio moves with temperature and season. Sandia National Laboratories notes that it typically runs lower in warm months, which in Florida covers much of the year. Compare the ratio with the system’s own design expectation for the same month, and track health with a performance index, which stays near 1 for a healthy system.
Does commercial solar monitoring replace site inspections?
No. Monitoring locates where output is being lost, and site inspection finds why. A Nextpower finding in kWh Analytics’ 2026 Solar Risk Assessment reported that 79 percent of high-risk PV connector failures showed no thermal signature at the time of inspection, which is why scheduled hands-on checks stay in the maintenance plan.
Why does my monitoring portal not match my utility bill?
The portal reports what the inverters produced, while the utility meter records net energy at the service point after the building’s own consumption. Under Florida’s net metering rule, on-site generation offsets consumption first, so the two numbers differ by design. Questions about how exports were credited go to the utility, which sets that treatment under its tariff.
When does an underperforming inverter need replacement instead of repair?
Replacement becomes the stronger option when the same unit faults repeatedly, parts or manufacturer support are no longer available, or the inverter is near the end of its service life. A single failed fan or board on a supported model can be a sound repair. The commercial solar inverter replacement guide sets out the repair-or-upgrade test in detail.
Why should an underperforming array be found in the data before it shows on the bill?
An underperforming array costs output every day the fault goes unread, and the utility bill reveals the loss weeks later, blended with the building’s consumption. The evidence above traces that loss to running inverters with open inputs, to thermal derating, to soiling films that build slowly in Florida humidity, and to shutdown alerts that wait for someone to act. Alarm-only monitoring catches only the last of these, and only when the alert reaches a person. Inverter-to-inverter comparison finds the single-unit faults, a performance index finds the shared ones, and a named owner with a dispatch clock turns either finding into a repair. On most arrays, that work starts with data the system already records, read against the right reference.
Written by The Technical Team at Advanced Green Technologies. Founded in 2007, Advanced Green Technologies is 100% employee-owned and was ranked first among Florida commercial and industrial EPC contractors on Solar Power World’s 2026 Top Solar Contractors list.
Request a monitoring review. Send your portal access and recent production data and we will review it against your system design. Request a free quote from Advanced Green Technologies or call (888) 735-2641.