Data collection & processing

The data calibration technology of irradiance estimated by the satellite-based CUBE relies on a 12-month on-site measurement campaign.

This consecutive 12 months duration, often conducted during the impact study phase, is ideal for collecting a sufficient data history. It allows capturing meteorological fluctuations specific to each season without constraining the project schedule.

Work conducted by CalibSun has demonstrated that calibration results are very robust over 12 months, and extending it to 24 months would yield only a slight improvement. Conducting a campaign of intermediate duration would bring no benefit and could even compromise performance, as some seasons would be observed twice and others only once.

The meteorological measurement cabin includes pyranometers for measuring irradiance, but not only that.

A series of weather sensors allows for 24/7 measurement of variables useful for project sizing, such as temperature, wind speed and direction, relative and absolute humidity, etc. Additionally, the cabin incorporates a soiling measurement sensor, allowing for refining assumptions made in the models.

It should be noted that the offer of sensors installed on CUBE is adaptable according to specific needs. In particular, it is possible to add optional sensors suitable for bifacial/agricultural projects (albedometers, diffuse measurement, etc.).

CalibSun is aware that the quality of measurements is paramount to ensure effective satellite data calibration that is faithful to the site's reality. That's why all sensors used are Class A, with special attention paid to pyranometers, which are calibrated before each campaign, mounted on a heating/ventilation system, and regularly cleaned during the measurement period (included in the offer).

CUBE is 100% autonomous, with oversized photovoltaic battery systems, and therefore suitable for off-grid installation at prospecting sites.

CalibSun includes in its CUBE offering operational monitoring of data collected on-site to trigger intervention if necessary.

This ensures maintaining the highest possible level of data availability and also monitoring any suspicion of sensor drift. In particular, the horizontality, calibration, and cleanliness of the pyranometers are regularly checked, with any beginning of drift corrected on-site.

CUBE clients have real-time access to the uploaded data to verify measurement availability.

Before performing calibration on the collected data, CalibSun ensures to discard any measurement point suspected of being faulty.

Multiple quality tests are applied, covering all measurement anomalies not visible to the naked eye. These tests include, for example, tests of extreme value limits, point shading by nearby horizon, outlier detection, etc.

The irradiance time series measured is then free from suspicious data, resulting in calibration of satellite data solely based on the reality of local meteorological phenomena, rather than on errors attributable to measurement.

The bias correction algorithm used in the CUBE solution is based on "quantile mapping," recognized as one of the most effective methods for meteorological data calibration.

CalibSun has made numerous methodological modifications to this approach to optimize performance for irradiance data, with a reduction in bias from satellite databases very close to perfection. Based on its published scientific studies presented in scientific committees (Read scientific paper), CalibSun is capable of affirming that reducing uncertainty in satellite databases decreases from 3% to 0.5%. This allows for a reliable P50 yield and a reduced P50/P90 distance.