New database for PV module fault detection

Scientists in Colombia have created a 10-year open dataset for PV panel fault detection. Data was acquired using an electronic load system, oscilloscope, and thermographic camera in outdoor environmental conditions.

Researchers from Colombia’s Metropolitan Technological Institute of Medellin (ITM) have aggregated an open dataset of PV performance under different fault conditions.

Available online, the set includes information from several panels that have been in outdoor operation for more than 10 years. “This dataset offers valuable insights into the performance of photovoltaic panels in real-world fault conditions, including discoloration, cracks, and shading,” the research group said. “This dataset can be used to develop and validate fault detection and classification algorithms. The combination of electrical and thermal data can also be used for building predictive models.”

The dataset consists of two types of PV modules. The first is Solar Plus Energy’s SP090P panel, which includes 36 cells and is rated for a 90 W capacity. Per its datasheet, the panel’s short circuit current is 5.41 A and its open circuit voltage is 22.5 V. The other included module is Erdm Solar’s Hybrytec-M5-30/12, with 36 cells and a capacity of 30 W. Its short circuit current is 1.93 A and its open circuit voltage is 21.6 V.

Three conditions were recorded for each panel type, namely normal condition, discoloration fault, and crack faults. Each was tested as clean, with dust accumulation, and it was also measured under partial shading. Key electrical parameters, such as short-circuit current, open-circuit voltage, and power output, were recorded alongside thermographic images that allow for thermal performance analysis. All panels were measured on May 17, 2024, from 10:00 am to 2:00 pm.

The online database includes folders with either raw data or processed data. The electrical section of the raw data includes voltage and current as well as the temperature and irradiation levels. Raw thermal data includes thermal imaging. The processed electrical data section includes I-V curve and power data, and the processed thermal data includes thermal information adjusted to identify the point of highest temperature in the panel area.

“This information can assist the creation of diagnostic instruments and enhance methods for increasing the longevity and effectiveness of solar panels by examining how frequent defects impact performance,” concluded the team.

The dataset was presented in “Dataset of Photovoltaic Panel Performance under Different Fault Conditions Cracks, Discoloration, and Shading Effects,” published in Data in Brief.