Generation of spatially dispersed irradiance time-series based on real cloud patterns

•Generation of Spatially Dispersed Irradiance Profiles (SDIPs) for photovoltaic arrays is proposed.•Real sky images are used to generate realistic irradiance profiles on inclined PV module surface.•Light transmission properties of different cloud types are considered during analyses.•It is demonstrated that the incident instantaneous irradiance values differs from point to point.•Daily irradiance sequences for different array observers are generated based on a set of sky images. Clouds, being complex components of the atmosphere, have significant effects on power generation by photovoltaic (PV) systems. For example, shadows caused by the cloud coverage over a geographically distributed PV power plant may cause significant fluctuations in power generation by leaving a number of PV panels unable to generate power and contribute to power generation by the plant at each time instant. Thus, investigation of the mentioned effects on PV power generation requires realistic spatial irradiance information. Such information should be evaluated based on the existing real cloud coverage and its light transmission characteristics. This also provides the opportunity to select appropriate coping strategies against the mentioned negative effects of partial shading on PV plant’s power output. This paper presents a modeling approach which generates Spatially Dispersed Irradiance Profiles (SDIPs) for PV arrays based on existing cloud patterns derived from local sky images taken at the application sites. The model gets the direct, diffuse and global irradiance values incident on a horizontal surface, which are primarily obtained utilizing a solar irradiance model (the Morf (2013) model), along with local sky images captured at the application sites and cloud transmittance values, as input data and yields site-specific Spatially Dispersed Irradiance Profiles (SDIPs) incident on the surface of inclined PV panels within PV application areas, as a result of process of the inputs. Utilization of local sky images and cloud transmittance values for different cloud types creates the opportunity for precise analysis of interactions of sunlight with the existing cloud type and hence, obtaining unique and site-specific irradiance profiles according to the existing cloud type and distribution in the sky. The model firstly detects the cloudy and clear-sky parts in the sky image and then instantly utilizes the most appropriate ellipse on the cloud layer associated with each solar panel thro