Frequently asked questions

You are right. Inverter faults and grid outages are SCADA's job, and by industry data the largest single loss item is the inverter. Even our own report refers inverter shutdowns to SCADA, not to us.

SCADA mostly does not show panel loss. When it does, it does not say where the loss is. On sites built with a DC/AC ratio above 1, the inverter clips at its AC limit in the summer months. During clipping hours a DC-side panel or string loss does not reach the output at all. Outside clipping hours and in winter, the loss of a single string stays below the variability of weather, temperature and irradiance. Days on which cool panels produce above expectation mask it further. Even with string monitoring, SCADA does not say which module, which diode and which cause are responsible. We say which row, which table and which module the loss is in.

Using a drone is the right call. The difference is not the drone, it is what happens after the flight.

A manual survey photographs the site and an operator reviews the frames by eye. The threshold is that operator's judgement, the call usually rests on a single frame, and no panel is given a lasting identity. What you end up with is a folder of images describing one day. For us the flight is only the first step. The frames are merged into one orthomosaic, AI finds every module one by one, the temperature difference is measured and normalised to irradiance to IEC TS 62446-3, and every finding is verified across several images. The real difference starts after that. Every panel is written to the platform under a permanent identity. You open your plant on a map, put the thermal and RGB layers over each other, click any panel and read its measured temperature difference, its class and its history. Findings are listed by priority and go to the field team as a work list. At the next inspection the same panel is compared under the same identity, so you see whether an anomaly has grown. A folder of images cannot give you that. We can also work as a module-level addition to your O&M provider's own report.

Asking what a new budget line returns is natural. That is why we state both lines openly.

The analysis fee is 200 USD + VAT per 2,500 panels. That is 0.08 USD per panel, with a minimum of 200 USD. Classified and prioritised findings, the inspection report and platform access are included in it. If we fly, a capture fee of 200-700 USD per day is added, set by the size and location of the site. It covers the cost of the flight and is fixed in the quote together with the number of days. If you fly your own RTK and thermal-equipped drone, you pay the analysis fee only. Prices exclude VAT. The TL equivalent is stamped at the day's rate on every quote. For comparison, by the industry average sites lose 5.77% of production a year. Most of the loss is concentrated in a small number of high-priority findings. That is why a prioritised report is valuable.

Soiling is a large item, and cleaning removes it. You may well be partly right about this. Which part you are right about can only be known by measuring.

A site has two separate kinds of loss, and only measurement separates them. One is the loss that cleaning removes, the other the permanent anomaly that cleaning does not touch. The effect of opaque dirt such as bird droppings is not confined to the surface it covers. It brings the bypass diode into play, and the resulting hot spot can turn into permanent damage over time. Our report totals the loss under these two headings separately. Soiling classes are read from the RGB image, electrical anomalies from the thermal image, and each takes its place on the map. You target the cleaning budget by the map and write the permanent anomalies into work orders. Your cleaning decision then also rests on measurement. Which row was cleaned or not cleaned, and why, is visible on the map.

Yes. If the hardware meets the requirement, the inspection works with your own flight.

There are two paths. We fly the mission, or your own team flies it with an RTK-equipped drone carrying a thermal camera. With your own flight there is no capture fee, only the analysis fee. The hardware requirement is this. RTK positioning is needed, and the route is not started without a FIXED solution. A radiometric thermal camera and an RGB camera are needed. The resolution must give at least 5×5 pixels per cell at 30 m. In both cases the data is collected to the same flight standard. The flight is made from 30 m, with lines crossing the rows and terrain follow at sloped sites. Before processing, the data passes the same acceptance check. That check covers coverage, row continuity, brightness uniformity and edge crispness. Data that misses the standard is not processed. The reliability of the result depends on whether the standard was followed, not on who flew.

An inflated finding list is a real problem in the industry. A type claimed from a single frame, exaggerated counts and a PDF that never becomes a work order are typical examples.

Our method is different. Every module is assessed from images taken from dozens of angles, not from a single frame. A finding enters the report only when it is measured consistently in more than one image. Every line is traceable back to the raw radiometric frame. What we deliver is not a finding list, but a prioritised work list.

Thermal measurement depends on the weather. The specification itself says so.

IEC TS 62446-3 requires thermal measurement in clear sun at 600 W/m² or more. That condition is met at midday on a clear day for most of the year. There is no season to wait for. Cloud, wind and humidity can move the flight day. Conditions are measured on the day and written into the report's inspection conditions section. We recommend one flight a year. Commissioning, the end of a warranty and a suspected drop in production each call for a flight of their own.

Within 5 business days of the data entering the system. Data that misses the standard is not processed.

The flight ends, the orthomosaic passes the four-point acceptance check, and processing starts. The report and the platform record are delivered within 5 business days of the data entering the system. You receive the orthomosaic mapping, a panel-level record, a finding list with types named and prioritised, a production loss analysis, an inspection report to the IEC TS 62446-3 standard, and platform access. 50 pages of a delivered 315-page report, finding cards included, can be downloaded on the site.

You keep your plant's panel record. That record is digital and grows with every inspection.

Every panel stays on the platform under its identity and position. Findings appear on the map, work orders are tied to the panel, and every inspection is added to the same record. A resolved anomaly closes on the same panel record at the next flight. Access is by account and role. There are separate roles for the field team, the operator and the investor.

The classification consists of the 17 classes of the IEC TS 62446-3 standard. Not all of them are seen in the same channel.

From the thermal image we detect single and multiple hot cells, hot spots, single and multiple diode anomalies, offline modules, string anomalies, warm junction boxes and reversed polarity. From the RGB image we read glass breakage, missing or detached modules, soiling, bird droppings and biological growth. Vegetation and structural shading are assessed from both channels together. Priority looks at safety as well as production loss. A burnt junction box and reversed polarity are P1 even where the loss is small. The full list, with real cuts from delivered inspections, is on the Services page.

There is no limit on site size. What grows is the number of flight days.

The analysis fee is the same everywhere, 200 USD + VAT per 2,500 panels. On large sites the flight is split over several days. If we fly, the capture fee is multiplied by the day count and written in the quote. Portfolios are followed on one screen on the platform. The date of the last flight, the number of open findings, the priority distribution and a comparison across sites appear on that screen. Because every flight is written onto the same panel identities, new, growing and resolved anomalies read against the same threshold across sites.