Siemens Aktiengesellschaft

Patrick Reeb

Joachim Bamberger

Ti-chiun Chang

Jan Ernst

Kuan-Chuan Peng

Aditi Roy

Deep learning is used to train a neural network for end-to-end prediction of short term (e.g., 20 minutes or less) solar irradiation based on camera images and metadata. The architecture of the neural network includes a recurrent network for temporal considerations. The images and metadata are input at different locations in the neural network. The resulting machine-learned neural network predicts solar irradiation based on camera images and metadata so that a solar plant and back-up power source may be controlled to minimize output power variation.

Solar irradiation prediction using deep learning with end-to-end training

A system and method for correcting short term irradiance prediction for control of a photovoltaic power station includes an irradiance data manager that generates an observed irradiance curve based on irradiance measurements received from an irradiation sensor, and a reference curve generation engine that generates a reference curve representative of irradiance values for a clear sky day. An irradiance prediction engine generates an irradiance prediction curve based on image segmentation and motion filtering of cloud pixels in sky images. A prediction correction engine corrects the irradiance prediction curve within a short term future time interval based on probabilistic analysis of time segmented intervals of the observed irradiance curve, the reference curve, and the irradiance prediction curve.

Automatic generation of reference curves for improved short term irradiation prediction in PV power generation

Systems and methods are provided for adapting images from different cameras so that a single trained classifier or an analyzer may be used. The classifier or analyzer operates on images that include a particular color distribution or characteristic. A generative network is used to adapt images from other cameras to have a similar color distribution or characteristic for use by the classifier or analyzer. A generative adversarial process is used to train the generative network.

Growing community of inventors

Adelsdorf, Germany

Patrick Reeb