It’s one thing to generate power using solar energy – it’s another to put that power to use in an actual power grid.

That’s one of the key problems photovoltaic (PV) experts have struggled with for years as they work to bring energy generated by solar panels further into the mainstream. One thing working against PV technology in this regard has been the fact that weather patterns make it difficult to predict energy output – a key ingredient to keeping grid flow balanced. Results from new research published in IEEE Journal of Photovoltaics, though, could help push PV technology a step closer to practical usage by more-accurately predicting solar panel temperature.

Right now our model is theoretical, but our results prove it’s ready to help integrate photovoltaics into the larger infrastructure of the grid, at least in Southeast Asia.

In order to integrate with the grid and keep the flow of electricity balanced, operators rely on short-term power output forecasting, which in the case of PV can be predicted by measuring weather variables in fast-time scales. Research shows solar panel temperature has a particularly strong influence on power output; thus, accurately predicting the panel temperature means a more accurate read on short-term power output forecasting.

Historically, however, operators have found it challenging to calculate this variable quickly and in transient conditions. This is because physical models currently used to calculate panel temperatures require many slow and difficult-to-obtain input parameters, which isn’t practical for the fast data-crunching demands of short-term forecasting. On the other hand, the existing empirical models for solar panel temperatures aren’t exact enough for reliable short-term power output forecasting. In addition, they’ve mainly been developed and tested in areas of high latitudes – such as Northern America and Europe – ruling out other climates.

To address these shortcomings, a team of researchers at the Dutch University of Twente in partnership with the Solar Energy Research Institute of Singapore (SERIS) developed a simple empirical model accounting for solar panel temperatures during transient weather conditions. The team took a standard empirical model called the Ross model and made it more precise by accounting for relative humidity, in addition to the ambient temperature, irradiance and wind speed which have been accounted for in previous models. They also incorporated the changes in these variables, rather than solely rely on their flat steady state.

The team then diligently tested its model on a minute-basis in two locations in Southeast Asia, Singapore and Jayapura, over several months. The team chose these locations because they’ve demonstrated particularly fast-growing, cost-effective market potential for PV systems.

“Solar panel temperatures are ever-changing, so the ability to simulate transient conditions adds a new dynamic layer to calculating output compared to the regular steady-state models,” said Hans Veldhuis, lead researcher. “Right now our model is theoretical, but our results prove it’s ready to help integrate photovoltaics into the larger infrastructure of the grid, at least in Southeast Asia.”

The team’s proposed model experienced a 46% improvement in accuracy compared to other empirical models, based on thorough calculations of root-mean-squared errors. As the average price for a completed PV system in the U.S. dropped 33 percent from 2011 to 2013, these results could help push PV technology yet one step closer to practical usage.

You can also find more articles about “Photovoltaics” in IEEE Xplore.