Improving Long-Term Back-of-Module Temperature Measurements
Inside this Article
Accurate and reliable back-of-module temperature measurements are essential for evaluating PV array performance. When you include other electrical and meteorological data, you can use back-of-module temperature measurements in concert with module temperature coefficients to monitor PV system performance, model predicted power output or assess warranty claims. (See “PV System Energy Performance Evaluations,” SolarPro magazine, October/November 2014.)
Many parameters drive back-of-module measurement accuracy and reliability, including sensor placement on the module, sensor technology, attachment method, and the balance of components in the data acquisition system. The better you understand the impacts of various measurement decisions—particularly, sensor type and attachment method—the more you can improve the accuracy and reliability of these measurements. Here I provide background on the topic and detail some best practices for measuring back-of-module temperature with improved confidence.
Measured vs. Actual Temperature
In considering the thermal environment of a photovoltaic cell, you are primarily interested in the temperature of the semiconductor (p-n junction). This is a difficult temperature to measure, since you cannot directly probe operating PV cells in fielded modules. As a proxy, you can use an open-circuit reference cell—which is a similarly packaged PV cell of the same technology—and extrapolate cell temperature from changes in open-circuit voltage. However, reference cells are built typically for measuring irradiance and are not readily available for measuring cell temperature.
As a result, you generally measure back-of-module temperature using traditional technologies, such as external temperature probes, and use these data as an approximation of the temperature at the semiconductor junction. Since multiple materials lie between the measurement probe and the p-n junction—including backsheet, encapsulant and semiconductor material—your back-of-module temperature measurements never perfectly reflect the temperature at the junction itself. Therefore, you must minimize the differential between the measured back-of-module temperature and the actual temperature at the semiconductor junction.
On one hand, the temperature coefficient of power for PV modules is a negative value, meaning that higher cell temperatures result in lower power output. On the other, poorly executed back-of-module temperature measurements usually result in measured temperature values that are lower than actual temperature values. If you inaccurately report the apparent back-of-module temperature as lower than it is in reality, you will overpredict the expected power output. As an example, the relative temperature coefficient of power for crystalline silicon modules is typically -0.45% per degree Celsius; therefore, if your measured back-of-module temperature is 7°C low, you will overpredict the expected dc power output by about 3.2%, which is a significant amount for large PV systems.
The sensors you are most likely to use for measuring back-of-module temperature include thermocouples, thermistors, resistive temperature detectors (RTDs) and infrared thermocouples. While each technology is theoretically capable of delivering reliable measurements over the lifetime of the device, I generally recommend Type T or Type E thin-film thermocouples for measuring back-of-module temperature. Within each device class, however, you must select among the available types or models to identify the specific sensors most suitable to the temperature range and environmental conditions that the fielded modules experience. Regardless of sensor technology, you must also pay attention to the sensor wire gauge or thickness. The data acquisition system (DAS) itself may also influence component specification.
Thermocouples. Thermocouples are constructed out of dissimilar metals or semiconductor materials, and they produce voltage in a predictable relation to temperature. While you may choose among many styles, thin-film and beaded thermocouples are most applicable for measuring back-of-module temperature. Thin-film thermocouples are formed from flattened or deposited metal traces on a plastic carrier. Beaded thermocouples are formed from twisted and soldered wire ends or by crimping the wires within a metal bead. As detailed in “Empirical Testing at NREL and Beyond,” test results indicate that thin-film thermocouples are typically more accurate than beaded thermocouples for back-of-module applications.