Raptor Maps, in collaboration with First Solar, the leading U.S. domestic manufacturer of solar panels, has made significant strides in the assessment of thin-film PV panel anomalies. By leveraging the firms’ joint expertise and resources, the partnership has resulted in enhanced efficiency for identifying glass crack anomalies in thin-film panels via drone-based aerial inspections. The results of the collaboration are a reduction in underperformance that leads to revenue loss and an increase in the longevity of American-made solar equipment.

Groundbreaking collaboration introduces advanced methods for identifying and addressing glass cracking anomalies, reducing underperformance, and enhancing solar asset management

Thin-film panel glass cracking–which is most commonly caused by mishandling, damage from on-site equipment, transportation, defects, or extreme weather–has historically been difficult to detect via aerial inspection due to the narrow width of top-sheet glass.* Excessive panel anomalies can lead to diminished power production and heightened safety risks; for example, if left unchecked, glass cracks may eventually allow water ingress, resulting in ground faults and potentially larger portions of the array offline.

In response to these challenges, Raptor Maps and First Solar engaged in comprehensive research during late spring 2023, focusing on developing inspection standards and refining image processing techniques that can replace on-the-ground inspections. The objective was to enhance top-sheet glass cracking identification accuracy and granularity in thin-film panels. These cracking anomalies are often very small, often only a few millimeters wide, requiring a more in-depth inspection standard beyond conventional drone inspections. To overcome this limitation, the Raptor Maps and First Solar research teams examined factors such as altitude, gimbal pitch, tracker angles, and environmental conditions. By establishing precise flight guidelines based on these factors, the study achieved consistent and improved rates of identifying all forms of top-sheet glass cracking.

Building from the close collaboration with First Solar, Raptor Maps leveraged the results from the in-field research to design and implement a novel approach to the identification of smaller instances of cracking, which allows for early detection and intervention before cracks propagate and significantly impact the panel's power production. Using machine learning, the new approach enables accurate and comprehensive identification of cracks. Historically, panel glass cracking identification and remediation has involved significant on-the-ground manpower. 

Via continued field research, Raptor Maps and First Solar are establishing thin-film-specific aerial inspection guidelines to guide the identification of thin-film-specific anomalies via software. These guidelines allow for the identification of defects observable in RGB imagery and thermal signatures specific to thin-film technology. By encompassing a broader range of anomalies, this comprehensive research further enhances the diagnostic capabilities of Raptor Maps' AI-powered software and facilitates the proactive maintenance of thin-film panels.

Through this collaboration, Raptor Maps and First Solar have improved PV panel anomaly detection, particularly in thin-film panels. By harnessing advanced techniques and leveraging extensive field research, the work dramatically improves the identification and mitigation of top-sheet glass cracking. The end results are enhanced operational efficiency, improved power output, and reduced safety risks within solar installations—all of which are a direct result of not needing to rely on field technicians surveying these types of anomalies. If you own and operate thin-film solar panels, reach out to us for more information about maximizing your power production: https://raptormaps.com/contact/

* Glass cracking is not specific to thin-film panels but can occur in any panel containing glass. For the purposes of this collaboration, thin-film panels were studied.