How to Detect Low-Energy Fractures in Photovoltaic Modules: A Technical Guide

Low-energy fractures are the most harmful type of fracture that can occur in a PV module since they are difficult to detect with the naked eye after just 24 hours as a result of their low force of impact and that they happen in a confined area on the module itself. This limits the area of damage for the initial fracture and does not necessarily cause additional fractures to propagate from this initial location. Because of their small size and confined area of effect, they are not usually easily identified during routine inspections. If these fractures go undetected and allowed to continue, they will reduce module energy production, increase series resistance and result in premature module failure. The purpose of this article is to provide information related to the different methods of identifying low-energy fractures in PV modules.

Photovoltaic (PV) modules have front glass that can develop low-energy fractures due to a variety of factors such as manufacturing stresses, vibration during transit, improper handling during installation, thermal cycles, and mechanical loads (e.g., hail hits or snow accumulation). Low-energy fractures look similar to large obvious damage, but do not immediately compromise the structural integrity of a PV module and will continue to expand if left untreated with normal operating conditions. This creates a significant challenge for asset owners/operators to locate their low-energy fractures since both standard RGB images and infrared thermography do not readily show these defects. Low-energy fractures therefore create a hidden liability in the context of large PV facilities.

The cornerstone of low-energy fracture detection is electroluminescence (EL) imaging, a non-destructive testing method widely recognized as the industry's gold standard for identifying internal defects in PV modules. EL imaging is formally standardized under IEC 60904-13, the international benchmark that specifies methods for EL imaging of photovoltaic modules.

EL imaging exploits the electroluminescent properties of semiconductor materials. When a PV module is forward-biased with a controlled direct current-typically 0.8 to 1.2 times the short-circuit current (Isc)-the PN junctions within the solar cells emit near-infrared photons (wavelengths of 800–1100 nm) through electron-hole recombination. A high-sensitivity camera-equipped with InGaAs or cooled Si-CCD sensors capable of detecting wavelengths from 900 to 1300 nm-captures these faint light signals and converts them into grayscale images. In a healthy cell, the EL image appears uniformly bright. In areas affected by defects-cracks, finger interruptions, fractures, or shunts-the abnormal recombination of charge carriers results in reduced luminescence, appearing as dark lines, dark spots, or localized dark regions.

For low-energy fractures specifically, EL imaging can reveal cracks as fine black lines or branching patterns that would otherwise remain invisible. The spatial resolution of modern EL systems can reach down to 10 μm, enabling the detection of even the most minute micro-cracks. The test is conducted in a light-controlled environment-typically a darkroom or with shading-to prevent ambient light interference. Exposure times typically range from 1 to 30 seconds, depending on camera sensitivity and module type.

Daylight EL Imaging: A Breakthrough for Field Inspections

Traditional EL testing has long required darkroom conditions, making it impractical for large-scale field inspections of operating PV plants. However, a breakthrough development emerged in mid-2026 when researchers from the Technical University of Denmark developed a novel daylight EL imaging method specifically designed to detect low-energy front glass cracks.

The innovation lies in transforming what were once considered challenges-sunlight and camera motion-into advantages. The method forward-biases the PV module with a modulated current, causing it to emit an EL signal. An InGaAs short-wave infrared (SWIR) camera then records multiple images while the camera is in slight motion, typically mounted on a drone. As the camera moves, cracked glass reflects sunlight differently from frame to frame. Software detects module corners, tracks and aligns the module across all frames, and applies fast Fourier transform (FFT) processing to extract the EL signal while reducing daylight noise. The reconstructed image reveals both conventional EL information-such as cell defects-and glass cracks, which become visible due to changing daylight reflections.

The team validated the technique through both controlled laboratory experiments and real drone inspections at the university's PV plant. Using a commercial drone equipped with a 640 × 512-pixel InGaAs camera, they found that the optimal imaging distance was 8–12 meters, with crack detection reliability decreasing beyond 15 meters.

Infrared Thermography: A Complementary Approach

Electroluminescent (EL) imaging detects cracks or fractures by looking at what is referred to as "light output." This method is non-invasive; electrical connections or wires do not impact how a module appears through an EL camera. Infrared (IR) thermograms depict another form of fracture detection using a different medium: heat signatures. Photovoltaic (PV) modules that are installed in sunny locations will maintain even temperatures across normal (or good) cells and also exhibit localized heating of defects due to their increased resistance to electrical current. The IR camera will capture these temperature differences, which may be in the range of several degrees to tens of degrees Celsius.

For low-energy fractures, infrared thermography can reveal the thermal consequences of a crack-increased series resistance leading to localized hot spots-even when the crack itself is not directly visible. Modern infrared detection systems have evolved from handheld devices to drone-mounted platforms, enabling rapid, non-contact inspection of entire PV plants. When combined with EL imaging, infrared thermography provides a more complete picture of module health.

Artificial Intelligence: Enhancing Detection Accuracy

The sheer volume of images generated by EL and infrared inspections-particularly in large utility-scale plants-has made manual analysis impractical. Traditional quality inspection required engineers to stare at screens, identifying defects by experience-a process that took at least one minute per module and was prone to fatigue, missed detections, and false judgments.

Artificial intelligence has transformed this landscape. Researchers have developed AI-based visual recognition methods that build on EL imaging technology. By constructing complete frameworks that include EL image acquisition, preprocessing, and advanced neural network models-such as improved Residual Networks (ResNet) with attention mechanisms-these systems achieve defect identification accuracy rates of 99.8% with a missed-detection rate of only 0.07%, processing a single image in just 8 seconds. This represents a dramatic improvement over both manual inspection and conventional algorithmic approaches.

An emerging approach for detecting fractures without interrupting module operation involves analyzing dynamic electrical response. This method monitors the electrical behavior of PV modules during normal operation and identifies anomalies indicative of cracking. Because it requires no system shutdown or specialized imaging equipment, it offers a practical solution for continuous health monitoring in solar energy systems.

Detection of low-energy fractures should be conducted in accordance with established standards. IEC 60904-13 provides the framework for EL imaging, while IEC 61730-2 addresses module safety qualification and includes provisions for breakage testing. IEC TS 62782 specifies methods for cyclic mechanical load testing, which can help identify modules susceptible to fracture under dynamic stress. For field inspections, the combination of EL imaging (particularly the new daylight method), infrared thermography, and AI-powered analysis represents the current state of the art.

The detection of low-energy fractures in photovoltaic (PV) modules is achieved through a multilayered strategy. The foundation of this strategy is the use of electroluminescent imaging (ELI), which has been advanced by the recent introduction of drone-based systems capable of carrying out large-scale field inspections during daylight hours. In addition, infrared thermography (IRT) provides complementary thermal data with which to assess the presence of low-energy fracture defects within the modules, and artificial intelligence (AI) technologies allow for rapid and accurate analysis at a large scale. With solar companies focusing on long-term performance over short-term growth, the ability to detect and address low-energy fractures will become increasingly more important to the long-term stability and bankability of PV assets.