Combination of Shadow and Irradiance Data to Create Irradiance maps

Kurzfassung

The exit from nuclear and fossil-fuel energy sources quires alternative energy sources to be tapped. Concentrating solar power (CSP) has proven to be an e�cient source of renewable power. In order to operate solar power plants, it is necessary to understand its e�ciency and energy yield in detail. If the direct normal irradiance (DNI) is only known for one or a few points in the solar �eld, errors can occur in the plant simulation and the plant control. The solar collectors, that provide the heat to the power block, cover approximately one square kilometer for 50 megawatt electrical turbine power. In this area, high variations of DNI can occur. For example, in parabolic trough fields four sub fields have their individual heat transfer fluid (HTF) flow control. Hence, the power plant control requires further information about inhomogeneous irradiance. In consideration of the fact, that the power plant operator wants to maximize the profit he needs to know, how the power plant will perform within the next 0-15 minutes. Forecasting in this range is called "nowcasting". For nowcasting applications it is beneficial to create irradiance maps with cameras, that observe the sky, to document the current irradiance for every coordinate of the observed area, since the operator is able to preact to short term di�erences in the DNI, which lower the power plant's energy outcome. There is a shadow camera system at Plataforma Solar de Almer��a (PSA), which is used to validate the nowcasting system. The system detects the cloud's shadows on the ground. It was extended from two to six cameras, that enable a 360� view around the installed tower power plant. Besides, the camera calibration was applied and older results were improved, so that an area of four square kilometers is evaluated. The cameras are recording series of two exposure times in 15 seconds intervals to ensure a high temporal resolution. There is also a sensor grid on the ground, measuring the global horizontal irradiance (GHI) and DNI. Its measured irradiance data was implemented to the shadow camera system. By using the shadow camera's sensor information, a relation between image information and incident irradiance was build. Therefore, it is possible to determine irradiance values for shaded regions, that are just de�ned by their image information. Plain view images (orthogonal images) are created for the evaluated area. By analyzing three time stamps, it is possible to detect shadow information. The shadow's structure is described by applying global thresholds. Also, the pixel difference is build between a shadowless and the investigated time stamp. Ultimately, these differences are used to define a DNI difference and generate irradiance maps for the evaluated area. Those resulting irradiance maps can be generated for each time stamp, if camera and irradiance data is given. It was possible to evaluate an entire day. The method's results are evaluated and its accuracy is discussed.