Investment in PV presents unique risks, as the amount of energy a PV system produces is as variable as the weather. Unfortunately, quantifying PV risk can be just as daunting as predicting the weather. Today, researchers have access to more research methods and more data than ever before. Now the search is on to identify the most accurate approaches to predicting PV plant output and quantifying the performance of existing plants.
These techniques to predict PV generation all play an important role in calculating risk for a particular project. Because reduced risk leads to less costly project financing, researchers have targeted nearly all major areas in PV modeling where uncertainty still looms, and are working diligently to make funding PV more palatable to all classes of investors.
In early May, Sandia National Laboratories in conjunction with the Electric Power Research Institute (EPRI) and the US Department of Energy (DOE) hosted a workshop focused on this topic. Attendees included model researchers, software companies that implement these models, and the independent engineers and PV plant developers that put the models into practice. Many independent engineering and plant development firms were represented because these models are particularly important in their work of assessing the potential energy output of an existing or planned PV system.
PV models, which are designed to best represent what happens in the real world, will always be couched in some uncertainty, as no model represents real world conditions exactly. As a result, many presentations focused on quantifying and reducing model uncertainty. From the quality of research presented, it was clear that the state-of-the-art in PV modeling has made significant strides in the past few years. The better a model can represent the physical reality of a PV system, the lower the uncertainty.
The conference was organized by topic area, starting with specific module-level model improvements, and moving on to actual implementation in project development and operations. For example, Dr. Cliff Hansen from Sandia National Labs demonstrated how outcomes from PV module simulations are often inconsistent due to indeterminate module level parameters. Hansen’s study not only showed how the user’s lack of knowledge about the model impacts results, but also revealed the lack of—and need for— standard practices for module-level simulations.
The start of the second day focused in an area particularly relevant to SolarAnywhere® , satellite derived irradiance data from Clean Power Research®. Different engineering consultants described methods for understanding and reflecting solar resource uncertainty as an input to PV models and the ultimate impact on energy assessment. Many presentations highlighted the benefits of using satellite-derived solar measurements by comparing this measurement source with high-accuracy, well-maintained ground based measurements.
The Sandia workshop provided an excellent opportunity for the industry to get up-to-speed on the latest modeling advancements. The ability to accurately quantify PV project risk will have a significant positive impact on the market through reduced costs and greater confidence in the ability of PV to deliver positive returns for investors. Forums like this workshop provide the setting where project developers and engineers can provide feedback to software developers and those implementing analytical models on how to effectively improve the accuracy and predictability of PV resource and power output models and tools.