Find which solar sites are losing money—and what to inspect first.

From telemetry to decisions

Solar portfolios create more signals than teams can review by hand. ArraySignal combines weather, configuration, and telemetry to answer one practical question: what needs attention first?

Portfolio prioritization

The portfolio view ranks sites by performance, data confidence, and modeled financial impact. Healthy assets stay visible. Data issues remain unknown. And the clearest production losses rise to the top.

Expected versus actual

At each site, ArraySignal compares measured AC power with calibrated, weather-adjusted expectations. Here, output pulls away from the expected curve through midday. And missing telemetry is never quietly treated as zero production.

Evidence-based diagnosis

Deterministic detectors check duration, irradiance, telemetry quality, and equipment behavior. The result is a probable fault backed by evidence—not a black-box claim, and never a confirmed diagnosis without field verification.

Energy and financial context

With sufficient data, the production deficit becomes estimated lost energy. A site-specific energy value adds financial context, while confidence and limitations stay attached. Unknown impact is never shown as zero.

Visible model confidence

Confidence stays visible alongside every estimate. ArraySignal scores weather quality, telemetry completeness, site configuration, timestamp integrity, and calibration history, so operators can see what supports the model, and where the evidence is limited.

Healthy counterfactual

For any selected period, the healthy counterfactual reuses the calibrated expected-production engine. It estimates how the site likely would have performed under healthy conditions, without automatically carrying active outages, soiling, or availability losses into the baseline.

Money leak and cost of waiting

The money-leak view turns eligible power deficits into estimated lost energy and revenue. For active incidents, a recent weather-adjusted daily loss rate supports seven, thirty, and ninety-day projections, clearly labeled as estimates, not promises.

Field dispatch economics

Dispatch economics compares the thirty-day avoidable-loss estimate with site-configured inspection and repair costs. The recommendation remains decision support: monitor remotely, investigate further, or consider a field visit when the expected value justifies it.

Historical fleet audit

Historical fleet audits apply the same pipeline across many sites: normalize telemetry, validate data quality, run expected production and fault detection, then rank estimated losses. One bad file produces a site-level error instead of stopping the portfolio.

Full historical data view

At site level, operators can move from the last seven days to all available history and keep the expected-versus-actual interval comparison in view. Missing weather or telemetry remains visibly unknown, never converted into invented production loss.

A defensible next action

Finally, ArraySignal combines probable cause, modeled impact, and inspection economics into a ranked action queue. Operators get a defensible next step, without promises of guaranteed recovery. ArraySignal: from portfolio signal to prioritized action.