Machine Learning Methods in Solar Photovoltaic Applications

Dear Colleagues,

The exponential growth of photovoltaic (PV) power generation has brought about new challenges to the grid integration, operation, and maintenance of PV power systems, where machine learning and automated data analysis methods are necessary to solve increasingly complex, multidisciplinary, and data-intensive problems.

First, the increase in PV generation combined with the intermittent nature of solar energy poses challenges to the stability and operation of the electric grid. Therefore, accurate intra-hour, intra-day, and day-ahead solar irradiation and PV system power forecasting are necessary to balance the energy supply and demand. Machine learning prediction models developed based on historical trends, meteorological measurements, numerical weather prediction models, satellite observations, cloud sky images, etc. have an increased potential to solve these problems.

Moreover, with the deployment of millions of solar PV panels—both in distributed systems as well as in large PV plants—efficient condition monitoring, fault detection, and localization is an increasingly difficult challenge. In small residential or commercial systems, the lack of local weather sensors makes monitoring the performance of the system difficult. Therefore, sensorless data-driven methods are necessary for detecting performance degradation or faults in the system. On the other hand, large PV power plants where large SCADA systems are deployed and continuously monitor the power generation face a data analysis bottleneck, where machine analysis and intelligent systems are required to detect anomalous operation, identify and localize faults, and recommend cost-effective maintenance actions.

Finally, PV system field inspection is transitioning to imaging-based diagnostic methods such as infrared (IR) thermography and electroluminescence (EL) imaging, performed by human operators or by drones. These inspection tools, although accurate, result in large image datasets that require automated image analysis, the detection and classification of failure patterns, and the prediction of failure severity on the solar panel’s power output.

This Special Issue focuses on new research and the application of machine learning and machine analysis methods for improving the operation and energy production of solar PV systems and supporting their widespread integration into the electrical grid. These issues include but are not limited to:

• Intra-hour, intra-day, and day-ahead solar resource forecasting;
• Intra-hour, intra-day, and day-ahead photovoltaic plant output forecasting;
• PV system fault detection, classification, and localization;
• Sensorless (weather) and data-driven condition monitoring;
• Early PV system degradation detection;
• PV system degradation prediction;
• Automated failure detection and classification from IR thermography and EL measurements;
• PV system power loss prediction from IR and EL measurements.

Dr. Sergiu V. Spataru
Guest Editor

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Applied Sciences is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• solar irradiation forecasting
• PV power forecasting
• fault detection
• fault classification
• anomaly detection
• condition monitoring
• data-driven methods
• image analysis
• infrared thermography
• electroluminescence imaging
• machine learning
• machine analysis
• prediction models
• clustering methods