Solar energy
Dr. Stefan Wilbert is team leader of the Solar Energy Meteorology team at the German Aerospace Centre (DLR) Institute for Solar Research. He is currently based in Spain, at the Plataforma Solar de Almería. This is a large solar power plant, where DLR runs a meteorological site for solar technologies. His presentation focused on these techniques: how they work, what meteorological information they require, and how we can measure these data at the site of the solar plant itself.
Meteorological data is a tool for cost reduction
Dr. Wilbert introduced various forms of solar energy technology. The most dominant type are the the well-known PV panels, that can be found on rooftops or in fields. They differ among each other in cell type, and can either have a fixed tilt or tracker. Concentrated Solar Power (CSP) and PV (CPV) make use of directed mirrors to focus solar radiation to a central point. Solar heating & cooling were also briefly mentioned. The efficiency of these techniques depends on site-specific information about irradiance, but also temperature, wind, humidity, dust, and so on. Measurements at the ground are used to make predictions of the current plant yield, but also as model input to calculate future yields. Meteorological data and models are vital to support operational decisions, leading to higher yields, and also to reduce costs and select optimal technologies for specific sites.
Of all measured parameters, irradiance is most important. Satellite data as well as local ground measurements are used to monitor the plant performance and model its future yield. Different types of radiometers a used. A so-called rotating shadowband radiometer (RSI) measures the global irradiance, and additionally features a rotating band that blocks the radiometer once every minute in order to measure diffuse radiation. Calibration of these local sensors is vital, since poor calibration can introduce large uncertainties, and thereby higher costs!
The spectrum of the incoming radiation is also analysed: all solar technologies and radiometers have unique spectral characteristics, and the radiation that reaches these panels is also a function of the atmospheric composition and path of the light. In order to account for these effects, we need radiative transfer models and measurements of aerosol concentration and properties to calibrate satellite observations.
Circumsolar radiation reaches the surface at a slightly different angle than disk radiation, which has implications for CSP plants.
Solar radiation is not only divided into direct and diffuse, but also in disk radiation (the sun itself) and circumsolar radiation (forward scattered radiation). Circumsolar radiation reaches the surface at a slightly different angle than disk radiation, which has implications for CSP plants. This radiation can be measured using a special camera, or sets of pyrheliometers. Alternatively, it can be derived from RSI the measurements. For tower plants, also beam attenuation, the loss of radiation intensity between the mirror and the tower, can be a relevant factor for design and operation.
Forecasting of the variability of irradiance is necessary for the plant control (backup), yield optimization, design of the plant as well as optimization of the grid (peaks of power). Forecasts are made in the very short term: 30 minutes up to hours. This very short-ranged ‘nowcasting’ is mainly done using sky-photos, taken at the ground from several cameras. Based on these photos, any clouds can be geolocated, a projection of their shadow is made, and this shadow map is combined with an irradiance map of the radiometers. So-called ‘shadowcams’ are cameras pointing at the ground, which are then used to create maps of the shadow over the plant, once again in combination with the radiometer data. These maps can be combined with satellite data to forecast plant yield for several hours (up to 6) ahead.
Not only irradiance, but also temperature, wind, moisture are important parameters for solar energy
Soiling (aggregation of dirt on the solar panels) can strongly inhibit plant yield. Meteorological data of dust, wind and precipitation are useful to estimate how often the panels need to be cleaned (which is not an easy task!). Reflectometers measure the amount of soiling on panels by comparing them to clean panels. Dust can also be measured directly by scatterometers or air filters. Soiling is just one example of processes that affect the durability of panels: temperature (mechanical stress), humidity (chemical reactions), wind (mechanical load & soiling) and sandstorms (abrasions) are a few other phenomena where meteorology provides important information. Accurate estimations of all these parameters can increase the lifetime of a plant.
Dr. Wilbert concluded his presentation with the message that, while irradiance is the main parameter, several others have to be considered and monitored by meteorological observations to increase efficiency. Meteorological data collection and application serve as a tool to reduce uncertainty and costs, to allow for optimal site selection and design, and optimal operation of the plant through forecasting and live data.
From the audience came the question whether cloud dynamics are involved in the nowcasting methods. As of yet they are not, which impeded the accuracy. Cloud dynamics, especially of shallow cumulus, can be very important on the 30 minute to 6 hour timescale of the energy forecasts. Measurements to estimate cloud optical thickness could be used to improve the accuracy of the irradiance forecasts. To the question of the influence of a solar plant on the local energy balance, dr. Wilbert replied there have been no real observations of changing weather caused by the plant, and that it might be similar to a paved urban surface in terms of energy balance. A final question is made about hail forecasts: are they relevant for solar plants? Typically, solar plants are located in lower latitudes where hail occurs only sporadically or not at all. In principle most panels should be able to withstand hail. However, in higher altitudes where hail occurs more frequently a hail event climatology might be a useful asset.