This article explains the thermal anomaly types that Sitemark identifies in your Solar Thermography data products. If you want more details on the Solar Thermography data products, check this support article.

In this article, we will list the thermal anomalies and their symptoms. We mention the causes per type that lead to the thermal anomaly. For more in-depth information about the causes, please go through this article. We also refer to the remedial action that might apply to the specific anomaly. Please go through this article for more in-depth information about the remedial actions.

Hot Spots indicate a defect at the cell level, in which one or several cells have a higher temperature than the neighboring ones.

Depending on the temperature difference between the cells, a Hot Spot may indicate a defect of varying levels of severity.

At Sitemark, we distinguish Hot Spots (only one cell appears to be impacted) and Multi Hot Spots (several cells of one PV module are unambiguously impacted).

Some real examples showing a Hot Spot in the thermal image.

The causes of a Hot Spot can be numerous. Some causes are reversible, while others are not. Thermal images usually aren’t sufficient to identify the cause of a Hot Spot, but thanks to visual (RGB) photos, we can identify the main cause of the Hot Spot or Multi Hot Spots.

The most common causes come from external elements that would cast a shadow on the PV module. As a cell gets covered by shade, it turns into a resistor and dissipates the energy produced by the other cells in the string. At Sitemark, we identify the cause according to the following categories:

Depending on the cause, suggested remedial actions will vary:

Overheated cells acting as a resistor will dissipate the energy generated by the other cells. Such cells will have an impact on the production of the full string of that PV module.

We generally assume that the losses will be proportional to the hot spot's delta temperature relative to the plant's production level. Be aware that one Hot Spot on a string is capable of causing a power production loss of up to 90% of the nominal production of one PV module.

Cells in a solar module are organized in substrings, ie. chains of cells connected in series. A typical PV module includes three substrings of 20 or 24 cells each (depending on the PV module type). To prevent shaded or broken cells from impacting the production of the whole string, these substrings can be bypassed thanks to bypass diodes.

These substrings will then be isolated and won’t produce any energy. As a result, they will have a higher temperature than the cells working normally in the same PV module. The defect can be seen in the thermal images, with one-third, two-thirds, or even the whole PV module warmer than the neighboring ones.

Some real examples showing a single bypassed substring in the thermal image.

Some real examples showing a double bypassed substring in the thermal image.

Bypassed substrings are a fail-safe mechanism of the PV modules. Generally speaking, this is expected behavior when the cause of the shading is visible, and the defect disappears when the cause of the shading is fixed.

However, bypassed substrings can be persistent in two main cases:

In most cases, bypassed substrings are reversible defects and will disappear as soon as the cause is fixed. If a bypassed substring is present despite no cause being visible, the most relevant fix would be to consider replacing the PV module. Depending on the age of the plant and the cause of the issue, manufacturers may cover the replacement of impacted PV modules under warranty.

For a single bypassed substring, one-third of the PV module is disconnected, and the power loss will be as high as one-third (33,33%) of the entire power production of one module.

For a double bypassed substring, two-thirds of the PV module is disconnected, and the power loss will be as high as two-thirds (66,66%) of the entire power production of one module.

Symptoms

PV modules with several cells at a high temperature can be classified as Multi Hot Spots. However, if these cells appear to be located in the same substring, the cause of the problem lies in the diode of that substring. One typically sees very high-temperature variations between the cells in one-third of the PV module, while the rest has a homogeneous temperature. This is what we call a single-diode issue.

Some real examples show a single-diode issue in the thermal image.

Similar to a single-diode issue, the multi-diode issue has more than one substring affected. The difference compared to a multi-hotspot issue is that the anomalies follow the line of the substrings and appear linear.

Some real examples show a multi-diode issue in the thermal image.

This defect is different from a single or a double bypassed substring. In a bypassed substring issue, the affected cells have the same temperature. In the case of a single or a multi-diode issue, all cells are hotter, but the thermographic images will show an irregular checkerboard pattern on that substring.

Causes

The diode designed to bypass the substring is damaged and will always let current through. Such damage can typically be caused by a very high current going through the PV module due to lightning, leading to a permanent malfunction of the diode.

Remedial Actions

The full PV module must be changed if replacing the diode is not an option.

Power Losses

Although complex to calculate accurately, the power losses can be estimated up to a third of the module’s nominal production per substring impacted.

PV modules show a pattern of hotspots starting from the edges, and this pattern starts at the negative pole of the string and goes a few modules to the positive pole. No visual marks.

This anomaly cannot be alone on a site, there must be more than one string showing this pattern on the site.

Two examples showing PID issue in some of the PV modules.

A voltage difference between the cells and the frame of the PV modules causes potential Induced Defects (PID). This difference may cause residual leakage current to flow through the cells and impact the yield significantly.

This effect is thought to be largely impacted by the following parameters:

Not all causes of PID issues can be acted upon, which leaves the operator with a limited choice.

The main remedial actions include:

Leak currents and overheating of cells will dissipate energy that the inverter would otherwise convert. A commonly accepted estimate in the scientific community is that power losses per PV module impacted may be as high as 30%. PID issues tend to get worse over time.

When there is not sufficient proof to mark an anomaly as PID but showing some symptoms then we mark it as Suspected PID, considering it is an early stage of a PID formation on the site. These doubtful situations are when there is no site overlay showing electrical connections of the modules, when the PID formed anomalies are not string level issues yet or not starting from the negative side but still showing a clear PID form.

Here is an example showing a Suspected PID issue in some of the PV modules.

Similar to PID, but with no short-term actions recommended. These defects can also indicate the PV modules' normal wear and tear due to the installation's age. Proper monitoring of the status of these PV modules is recommended to observe how the situation evolves.

Temperature differences undoubtedly indicate power losses, but it is assumed that such PV modules will cause lower power losses than actual mature PID.

One group of PV modules is warmer than another comparable group of PV modules by a few degrees, but the temperature is uniform across the impacted group.

Here is an example showing a string (open circuit) issue in the thermal image.

The full string is disconnected and gets warmer because the solar irradiance isn’t converted into power. The PV modules in the affected string absorb solar irradiance, which is why the surface of the PV modules in this string is hotter than the neighboring string. This may occur due to several issues, such as:

Start with a thorough check of the string connections, and these remedial actions should be simple. Since these issues cause extreme power losses and are often easy to fix, they should be treated with the highest priority.

When an entire string of PV modules is disconnected, these PV modules are not producing any power. This means that the power loss is 100%. for as long as the string issue has existed.

One group of contiguous panels is showing a patchwork pattern (panels with multi hotspots), usually the whole string.

Here is an example showing a string (revered polarity) issue in the thermal image.

The cause of this anomaly is difficult to be defined without necessary measurements on site. This is usually caused by short circuit, low voltage, reverse current or a mismatch on the string’s working conditions.

Start with a thorough check of the string connections, and these remedial actions should be simple. Since these issues cause extreme power losses and are often easy to fix, they should be treated with the highest priority.

When an entire string of PV modules is disconnected, these PV modules are not producing any power. This means that the power loss is 100%. for as long as the string issue has existed.

The whole combiner box looks precisely the same as an open circuit string.

When an entire combiner box is not working, the power loss is 100%. for as long as the combiner box issue has existed.

The whole inverter looks precisely the same as an open circuit string.

Causes can be;

When an entire inverter is not working, the power loss is 100%. for as long as the inverter issue has existed.

One PV module is warmer than the other PV modules by a few degrees, but the temperature is uniform across the impacted module.

Here is an example showing an open module issue in the thermal image.

The complete PV module is disconnected and gets warmer because the solar irradiance isn’t converted into power. The PV module absorbs solar irradiance, which is why the PV module's surface is hotter than the neighboring module. This may occur due to several issues, such as:

When an entire PV module is disconnected, it means that this module is not producing any power. This means that the power loss is 100%.

It is expected that junction boxes would slightly warm up because of the high current density it undergoes. Still, a high temperature may indicate issues in the connections that will cause power losses or even a fire hazard.

Any heat generated in the junction box is thus dissipated through the back sheet, wafer, and glass layers before it can get measured. This information is what we use to try and assess whether a junction box has a problem or not.

Our assessment of the health of a junction box is thus mainly relative, comparing the brightness of junction boxes compared to its neighboring ones. This is based on the assumption that if most of the PV modules show a warm junction box, they are not likely to be faulty. The temperature increase can be due to a high irradiance or a specific PV module design.

The example below on the left shows one junction box warmer than the surrounding ones indicating a potential problem. The example below on the right shows PV modules in which the temperature of the junction boxes looks similar across all PV modules.

Be aware that the location of the junction box may differ depending on the type of PV module.

The thin film cross-cell and multi-cross-cell anomalies resemble the hotspot and multi-hotspot anomalies seen on mono/polycrystalline panels. The difference, however, between mono/polycrystalline and thin film panels is that the cells in a thin film panel are long and thin, as shown in the schematic below (the red rectangle identifies one cell).

Single Cross Cell Multi Cross Cell

As a consequence of the narrow cells, some anomalies that would be classified as hot spots and multi-hotspots on mono/polycrystalline panels do not appear as single-point hotspots in the thermal imagery for thin film panels. The anomaly appears elongated and can appear to cross several adjacent cells and sometimes the complete width of the panel, as shown in the thermal examples above.

The cause of the thin film cross-cell anomalies could be related to a potential physical internal problem with the panel itself or could be due to the presence of broken glass, which is not visible in the RGB imagery.

Based on the severity of the anomaly, it is recommended to either monitor the anomaly for low-temperature delta anomalies or anomalies with a severity of Critical; then, it is recommended to repair the panel.

Although complex to calculate accurately, due to the extent of the panel covered by the anomaly, power losses can sometimes be up to a third of the module’s nominal production.

Some panels are fully or partially not visible due to physical obstructions such as trees or other lower vegetation grown close or between panels. Since the panels are partially or fully not visible, it is not possible to make analyse them. Therefore, this is not an anomaly but just a warning that some part of the site cannot be analysed

Here are some examples showing obstruction issue in the thermal and visual imagery.

Cause can be any physical obstructions such as trees, other lower vegetation grown close or between panels or overhanging objects.

Remove the obstruction

Since there is no visibility on the panels/some part of the panel, the power loss can not be estimated in this case.

There aren’t hotspots, but cold spots due to shading of objects such as trees, other small vegetation, electric poles, fences etc. In these cases, it is expected that these shadings are going to turn into hotspots in time if the shading stays, however, most probably this is a temporary shading and goes away or changes its position as the position of the sun changes.

Here are some examples showing shadowing issue in the thermal and visual imagery.

Cause can be any objects such as trees, other small vegetation, electric poles, fences etc.

Remove the material causing shadowing

Since this is most probably a temporary shading and goes away or changes its position as the position of the sun changes, a power loss estimation can not be made in this stage, however, if the shading stays and causes overheating on the panels then it must be analysed again.

The tracker doesn't follow the sun, it is quite clear on both thermal and visual imagery that the tracker looks a different direction then the others. When the malfunctioning tracker is on and has current running, it usually shows a different thermal pattern since it receives a different amount of irradiance then the ones working properly. In some cases this may cause mismatch between the parallel connected strings, therefore, causes issues with those parallel connected string even though they work properly.

Any mechanical problems with the tracker mechanism can cause a tracker malfunction.

Fix the tracker mechanism

The data is missing on a section of the site where there are operational modules to be inspected.

There might be different reasons for missing data, some of which are;

If the data is acquirable flying again is the only remedial action unless it is a data upload issue.

The module is missing on imagery but exists on the site overlay/design file.

This anomaly type is used only on non-standard broken glass inspections which are visual inspection with only visual (RGB) data that are done per specific customer request. There is no thermal data and the visual data has usually higher resolution then standard data. The anomaly has three certainty levels, therefore, three different anomaly types;

Here are some examples showing broken glass issues in the visual imagery. These examples may differ based on module type.

This anomaly type is used for issues detected on visual data but cause no thermal anomaly yet. Normally we do not do a separate visual analysis, however, when something that is considered to be worth to be specified is detected on visual data is encountered, it is labelled as "visual" anomaly and the cause will be added.

These visual anomalies may be caused by, broken glass, soiling, vegetation, dropping, removable objects or other objects which are visible on RGB data but not on thermal data yet.

Removing the cause should fix the problem.

Indescribable distinct cold spots on modules on thermal data, that are considered to be worth to be specified.