Abstract
The atmospheric conditions existing where concentrated solar power plants (CSP) are installed need to be carefully studied. A very important reason for this is because the presence of clouds causes drops in electricity generated from solar energy. Therefore, forecasting the cloud displacement trajectory in real time is one of the functions and tools that CSP operators must develop for plant optimization, and to anticipate drops in solar irradiance. For short forecast of cloud movement (10 min) is enough with describe the cloud advection while for longer forecast (over 15 min), it is necessary to predict both advection and cloud changes. In this paper, we present a model that predict only the cloud advection displacement trajectory for different sky conditions and cloud types at the pixel level, using images obtained from a sky camera, as well as mathematical methods and the Lucas-Kanade method to measure optical flow. In the short term, up to 10 min the future position of the cloud front is predicted with 92% certainty while for 25–30 min, the best predicted precision was 82%.
Highlights
The climate of Mexico City is tropical even though the city lies at an altitude of 2250 m above mean sea level [1]
We present the results for forecasting cloud positions based on their displacement trajectory
To obtain a cloud movement forecast with sufficient time in advance, we used images in which the clouds come from the horizon following a defined trajectory in accordance with the sky camera perspective
Summary
The climate of Mexico City is tropical even though the city lies at an altitude of 2250 m above mean sea level [1]. Along with its typical meteorological characteristics, means that convective clouds can develop with warm bases and cold tops, where ice particles form and develop during the early stages of precipitation [2]. In places where it is planned to convert solar energy into electrical energy, it is necessary to forecast the solar irradiance resource for operational optimization. Short-term forecasting is needed for operational planning, switching sources, programming back up and short-term power purchase, as well as planning for reserve usage and peak machine load [3]
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call