Clouds have a significant impact on the amount of shortwave radiation that is reflected, transmitted and absorbed by the atmosphere. On a local scale they are the decisive factor for short-term variability of solar irradiance reaching the ground. Hence, they are strongly influencing power outputs in solar energy applications.
With an increased share of renewable energies, particularly solar, information about current and future states of solar irradiance is therefore valuable and can lead to cost reduction and risk minimization for the operation of solar power plants and electrical grids. To determine the effects of clouds on current and short-term solar irradiance in high resolution, ground-based imaging using so-called All-Sky-Imagers has become a widespread field of research.
However, precisely localizing and classifying clouds in the sky can be a challenging task, for instance, due to variable form and shape of clouds, oversaturated image areas in vicinity to the sun, large distances to the camera, and overlaying cloud layers. As a result, current methods for cloud detection and classification often show high error rates.
An alternative approach that could lead to better short-term solar irradiance forecasts (nowcasts) would be to estimate the model’s uncertainty and provide probabilities of detected cloud types and their location. So far probabilistic cloud detection has not been addressed in the literature and the impact on solar irradiance nowcasts needs to be investigated.
In the context of this work, a probabilistic cloud detection model needs to be implemented and compared to an existing deterministic counterpart. The focus will be especially on the development of deep learning models that have proven to perform best on many computer vision tasks, like classification, object detection and segmentation.
Look forward to a fulfilling job with an employer who appreciates your commitment and supports your personal and professional development. Our unique infrastructure offers you a working environment in which you have unparalleled scope to develop your creative ideas and accomplish your professional objectives.
Our human resources policy places great value on a healthy family and work-life-balance as well as equal opportunities for persons of all genders (f/m/x). Individuals with disabilities will be given preferential consideration in the event their qualifications are equivalent to those of other candidates.
For a detailed job description, have a look on the job portal of DLR: https://www.dlr.de/dlr/jobs/en/desktopdefault.aspx/tabid-10596/1003_read-49708/