Trackable independently establishes what your solar fleet actually did, outage by outage, day by day, reconciled to the revenue meter. Then it puts that on the record, in findings your engineer can check.
Most of what this site lost to outages never reached the owner's service ledger, and long outages only ever ended in daylight. We find it, show how each outage began, and name the records that settle who pays.
Performance ratios inherit whatever the sensors say. A mislabeled or drifting sensor rewrites a site's reputation for years, and nothing flags it.
Production stops, and no record says why or who pays. The loss quietly becomes yours.
At refinancing and exit, an unverifiable operating history is priced against you. The caveat costs more than the kilowatt-hours.
Verdicts from the data your fleet already produces: which records can be trusted, every outage timed and classified, with the records that settle its origin named, and a disposition per site, with no hardware and no site visits.
It starts with the site’s own records: a year of inverter alarms, unit-against-unit comparisons and sensor checks, before any hardware ships. Where those leave a question, our edge instrument measures each inverter's AC output at 960 samples per second, reconciled to your revenue meter, daily. Where the array itself is in doubt, we name the strings to test. The truth of a site, established once, at instrument grade.
With or without an instrument on site. From the data your fleet already produces, every inverter is checked every day: alarms decoded, each unit measured against its neighbors, and anything new reported within a day. Running now at our pilot site, it flags anything urgent by email each morning. With an instrument, it adds an independent measurement that captures how a fault develops at 960 samples per second and reconciles to the revenue meter. Forensic reports follow significant events. Every measurement is written once and locked for seven years, so the performance number you hand a lender, an auditor, or a buyer rests on measurements kept unaltered, with every finding traceable to them, and comes without the data-quality caveat.
Every inverter, every morning. Anything urgent arrives by email the same day.
What each site did that month, reconciled to the revenue meter, with every outage accounted for.
The state of every site in one place, each number showing its source and when it was last checked.
Questions in plain English, by chat or text, answered from the record.
Monitoring platforms are starting to add chat assistants. Those answer from the platform's own averages and its own idea of expected output, the very data that needs checking. Ask the fleet answers from independent records reconciled to the revenue meter, on any monitoring platform. Every claim cites the record behind it, and every answer says what the data cannot show. Not built yet. The exchange below uses a made-up fleet.
Site 07 was offline for 2 hours 45 minutes on June 14, from 14:02 to 16:47 EDT. Every inverter tripped together on grid protection, and none on an equipment fault.
All six inverters went offline in the same polling window E-114, each with a grid-protection code at its first reading and no equipment code E-115. That shows the site lost its grid connection. It does not show whether the utility's equipment or the site's own breaker opened. The revenue meter recorded no export for the whole window E-116.
The monitoring platform logged no downtime that day E-117. Against the irradiance at the time E-118, lost generation is 2.1 MWh, upper bound 2.4 MWh. None of it is booked anywhere yet.
Site 07 offline Jun 14, 14:02 to 16:47 EDT. All units tripped on grid protection (classified); origin open pending the utility log. Meter confirms no export (verified). Not in the site's downtime log. Lost 2.1 MWh, upper bound 2.4.
Ref: E-114 to E-119
trackable.systems/r/07-0614
Your engineer can verify every claim. Your buyer's engineer can reproduce every result. That's not a courtesy. It's the product. Independence is what a record is worth.
The Fleet Health Check reads the monitoring data you already have and returns every site tiered, with reasoning for where verification would matter most and why. One week. Read-only access. Yours to keep. It is a free first look; Level 1, the forensic screen, is the full version.
We ask for nothing but a conversation. Your data stays yours, always.