Learnings from Thermography of over 12 GWs of Solar PV Plants

By Jeff Sykes on

Solar PV plants consist of hundreds of thousands of discrete components exposed to various stresses throughout the asset lifecycle. Many of these stresses can lead to defects that manifest as hotspots and impact the plant generation. Timely identification and resolution of these issues reduce generation losses and safeguard equipment. Drone-based infrared thermography coupled with Machine Learning analytics improves detection accuracy, and reduces the time to detect and rectify hotspots.

Thermography

From drone thermography studies of over 12GW of solar PV plants (over 30 million modules) across five continents, we have learned that:

  1. The median prevalence of hotspots was 1.4% of the total installed modules

The median prevalence of hotspots was 1.4% of the installed modules, while the average incidence was 2.1%.

  1. Best plants had <0.1% hotspot prevalence compared to 30% for the worst plants

In the utility-scale solar PV plants we surveyed, 0.05% to 30% of the total modules installed were affected by hotspots.

  1. Hotspot induced energy loss estimates range from 0.5%-3% of total generation

While most plants met PR Warranty levels, theoretical modeling of losses indicates that hotspots can cause energy yield losses ranging between 0.5-3%.

  1. Median hotspot temperatures found to be 10-15oC on an STC[1] adjusted basis

The typical median temperature gradient of hotspots was 4-7oC on a nominal basis and 10-15oC on an STC adjusted basis.

  1. Single-cell hotspots typically have the highest temperature gradients

Single-cell hotspots tend to have the highest temperature gradients. These hotspots are typically caused by shadowing due to vegetation, bird droppings, a thick dust layer, or malfunctioning cells with microcracks or defects.

  1. Rooftop solar fares poorly compared to utility-scale plants

Rooftop solar plants are worse off compared to utility-scale solar plants. The median prevalence of hotspots was 3.6% compared to 1.4% for utility-scale plants.

  1. There is no correlation between the size of the plant and hotspot prevalence

We did not find a correlation between the plant’s size and the prevalence of hotspots. There are plants >100 MW in size with less than 0.1% capacity impacted by hotspots.

  1. Drone thermography returns investment as early as 3 months

Depending on project size, yield improvement, and PPA rate, a thermal scan’s payback tends to be in the range of 3-6 months.

Thermography

Given the ease of thermal data collection and analysis with drones, we recommend the following:

  • Thermal drones should become an integral part of the solar PV plant O&M toolkit.
  • At a minimum, a once-a-year 100% thermal scan is desirable.
  • More frequent scans with in-house drone scan capability are likely to give maximum benefit for large solar plants.
  • Surveys should be conducted after panel cleaning to avoid noise from dirt-related hotspots.
  • Use software integrated with robust AI/ML-based analytics to enable rapid diagnosis and rectification of hotspots.
  • The use of software with ticketing and workflow functionality enables rapid identification and rectification of hotspots on site.

 

Are you conducting or planning to conduct thermography for your solar PV plants? Please contact us to discuss your requirements and propose a cost-efficient plan for thermography that helps you improve plant performance and health.

 

[1] Standard Test Conditions: normalized to 1000 W/m2 irradiance

Compare quotes from up to 7 pre-vetted installers in your area now.

Since 2008 our knowledge and sophisticated software has allowed over 300,000 Australian households and businesses to make a well-informed choice on their solar & battery installer.
Jeff Sykes
Latest posts by Jeff Sykes (see all)

Related Posts