Losses are found in days, not years
A dead string on a 400 kW array can hide inside monthly totals. It cannot hide inside a daily expected-versus-actual comparison.
Solution
Monitoring and O&M
An array is a twenty-five year asset with no moving parts and a quiet failure mode. Strings drop out, fuses blow, optimisers die and nobody notices, because the building still has power and the bill still arrives.
- Typical range
- All systems, in or out of our build
- Built for
- Any commercial array past commissioning
Overview
What this actually involves
Monitoring only earns its keep at the right resolution. Whole-system output tells you nothing until the loss is large. String-level or module-level telemetry tells you which combiner, which string, which module, and it tells you this week rather than at the annual true-up.
We model expected output for every system we monitor: irradiance and temperature corrected, degradation applied, soiling accounted for. Actual is compared against expected daily. When the gap crosses the alert threshold, a ticket opens automatically with the string identified.
Soiling is the Valley's defining maintenance issue. Between the dust, the agricultural traffic and a dry season that runs from May to October, an unwashed array in Kern County can lose a meaningful share of its output before anyone reacts. Washing is scheduled against measured soiling loss, not against the calendar.
Then there is the paperwork nobody wants. Module and inverter warranty claims require serial numbers, commissioning data, fault logs and a defensible production history. We keep that file from day one so a claim in year nine is an email, not an archaeology project.
Scope of supply
What is included
- String or module-level telemetry with revenue-grade metering at the point of interconnection
- Expected-versus-actual production model, irradiance and temperature corrected, checked daily
- Automatic alerting with a named threshold and a response time written into the agreement
- Scheduled module washing, timed against measured soiling loss rather than a fixed calendar
- Annual thermal imaging of modules, combiners and connections to catch hot spots early
- IV curve tracing on a rolling sample of strings, benchmarked against commissioning data
- Torque checks on racking and electrical terminations, plus vegetation control on ground-mounts
- Warranty administration: serial registry, fault logs, claims filed and chased to resolution
Sequence
How it runs
- 01
Baseline
Commissioning data becomes the reference. Every later measurement is compared against it.
- 02
Daily comparison
Weather-corrected expected output against metered actual, per inverter and per string.
- 03
Alert and triage
Threshold crossed, ticket opened, string identified, and a remote diagnosis before anyone drives out.
- 04
Scheduled service
Washing, thermal scan, IV traces and torque checks on a published annual calendar.
- 05
Annual report
Measured kWh, performance ratio, degradation trend against warranty, faults and their resolution.
Why it works
What you get out of it
Washing when it pays, not on a schedule
Soiling loss is measured, so washes happen when the recovered kWh exceeds the cost of the wash.
Warranty claims that succeed
Serial numbers, commissioning traces and fault logs on file mean a manufacturer claim has the evidence it needs.
A performance history for the building
If you sell or refinance, a documented production record is an asset. An undocumented array is a question mark.
What moves the price
Cost and schedule drivers
| Factor | Effect on the project |
|---|---|
| Monitoring resolution | System-level, string-level or module-level changes both the cost and how small a fault you can see |
| Soiling rate | Dust, agricultural traffic and distance to unpaved roads set the wash frequency |
| Site access | Roof access, fall protection and gate procedures affect every service visit |
| Array age | Older systems need more IV tracing and more connector inspection per visit |
| Response time committed | A next-business-day response costs more than a five-day one and is worth it on revenue-critical sites |
Often paired with
Related solutions
Answers
Monitoring and O&M: common questions
Not covered here? Our engineers answer directly, not through a call centre.
Ask a question(661) 555-0183It depends entirely on dust exposure and rainfall, which is why we measure rather than quote a number. What we can say is that in a Valley dry season that runs May to October, soiling is normally the largest single recoverable loss on an unmaintained array.
Yes. We commission a baseline, register the equipment, and take over alerting and service. Existing monitoring hardware is used where it is capable enough; where it only reports whole-system output we usually recommend an upgrade.
That the system will produce at least a stated percentage of the modelled annual kWh, weather adjusted, with a remedy if it does not. The weather adjustment is the part that matters: without it, a cloudy year looks like a fault.
Modules degrade slowly and predictably. Manufacturers publish a first-year figure and a subsequent annual rate, and warrant a retained output at year 25. We track measured degradation against the warranted curve so a genuine claim is provable.
Bring us the bill before you bring us the roof
Every project starts with twelve months of interval data and a load shape. The array comes after that, sized to what you actually use.
CSLB #1071482 (sample). Prevailing wage on public works. Engineers answer, not a call centre.