The energy available in Micro Grid (MG) that is powered by solar energy is tightly related to the weather conditions in the moment of generation. Very short-term forecast of solar irradiance provides the MG with the capability of automatically controlling the dispatch of energy. We propose to achieve this using a data acquisition systems (DAQ) that simultaneously records sky imaging and Global Solar Irradiance (GSI) measurements, with the objective of extracting features from clouds and use them to forecast the power produced by a Photovoltaic (PV) system. The DAQ system is nicknamed as the Girasol Machine (Girasol means Sunflower in Spanish). The sky imaging system consists of a longwave infrared (IR) camera and a visible (VI) light camera with a fisheye lens attached to it. The cameras are installed inside a weatherproof enclosure that it is mounted on an outdoor tracker. The tracker updates its pan an tilt every second using a solar position algorithm to maintain the Sun in the center of the IR and VI images. A pyranometer is situated on a horizontal support next to the DAQ system to measure GSI. The dataset, composed of IR images, VI images, GSI measurements, and the Sun's positions, has been tagged with timestamps.

We built a system composed of VI and IR solar radiation cameras, a tracking system, anda pyranometer, to collect the data. A weatherproof enclosure contains the two USB cameras,and it is mounted on top of two servomotors. The cameras are connected to a motherboard thatruns a solar tracking algorithm. This motherboard is placed inside of a different weatherproofenclosure together with a router, a servomotor control unit, a data storage system, plus theirrespective power supplies. The enclosure’s degree of protection is IP66. This is an internationalstandard utilized for electronic equipment that provides protection against dust and water.

The tracking system and DAQ software is currently operative. The software was programmed in Python 2. The system is placed on the top roof area of the Mechanical Engineer-ing building at UNM central campus (35.0821,−106.6259). The DAQ sessions can be visuallymonitored through a webpage. All devices were interconnected via a LAN network built by DHCP server.

The repository has available recordings of the solar cycle from 244 days of 3 years. The total amount of data is 110Gb. The sampling interval of the cameras is 15 seconds and the observation period is when the Sun's elevation angles is higher than 15°. There are approximately 1,200 to 2,400 captures per day from each camera depending on the day of the year.

• VI images are 16 bits with resolution 450 x 450, intensity channel only. Approx. 240KB per frame. Between 200 MB to 400 MB per day depending on the amount of images in the directory. Images are saved in .png compress-less format in the directory   */visible. The image are named by the UNIX time in seconds.    
• IR images are 16 bits with resolution 60 x 80. Approx. 8KB per frame. 20MB per day roughly constant. Images are saved in .png compress-less format in the directory */infrared. The image are named by the UNIX time in seconds.
• The pyranometer is sampled from 4 to 6 times per seconds. The measures are saved in the directory */pyranometer in .csv files named with their date (yyyy_mm_dd), approx. 4,500KB to 7,500KB per file. The files contain unix time in the first column and GSI in W/m² in the second column. 
• The Sun position files are generated out of the time recorded in the pyranometer files. The positions are saved in the directory */sun_position in .csv files named with their date (yyyy_mm_dd), approx. 6,500KB to 11,500KB per file. The files contain the unix time in the first column, elevation angle in the second column, and azimuth angle in the third column, all of them computed with full precision unix time.
• The weather station sample interval is 10 minutes. Linear interpolation was applied to match the sampling interval of the pyranomter. The weather station files are saved in the directory */weather_station in .csv files named with their date (yyyy_mm_dd), approx 14.3MB top 25.4MB per file. The files are organized by columns which contain from left to right: unix time, temperature in °C, dew point in °C, atmospheric pressure in mmHg, wind direction in radians, wind velocity in mile/s and relative humidity in %.