Forecasting Solar Radiation

Solar resource forecasting is very important for the operation and management of solar power plants. Solar radiation is highly variable because it is driven mainly by synoptic and local weather patterns. This high variability presents challenges to meeting power production and demand curves, notably in the case of photovoltaic (PV) power plants, which have little or no storage capacity. For concentrating solar power (CSP) plants, variability issues are partially mitigated by the thermal inertia of the plant, including its heat transfer fluid, heat exchangers, turbines and, potentially, coupling with a heat storage facility; however, temporally and spatially varying irradiance introduces thermal stress in critical system components and plant management issues that can result in the degradation of the overall system’s performance and reduction of the plant’s lifetime. The variability can also result in lower plant efficiencies compared to operation in stable conditions because optimally operating the plant is more challenging. For PV power plants that have battery storage, forecasts are helpful to schedule the charging process of the batteries at the most appropriate time, optimize the fractions of electricity delivered and stored at any instant, and thus avoid the loss of usable energy. Solar radiation forecasting anticipates the solar radiation transients and the power production of solar energy systems, allowing for the setup of contingency mechanisms to mitigate any deviation from the required production. With the expected integration of large shares of solar power, reliable predictions of solar power production are becoming increasingly important as a basis for efficient management and operation strategies as well as for solar energy trading. Today, solar power prediction systems are an essential part of electric grid management in countries that have substantial shares of solar power generation, among which Germany is a paradigmatic case. For example, in 2016 Germany had an installed PV power capacity of more than 40 GWpeak, supplying more than 40% of the total load on sunny summer days at noon. In this context, and according to the German Renewable Energy Sources Act (a set of laws aimed at promoting renewable energies in Germany), transmission system operators are in charge of marketing and balancing the overall fluctuating PV power feed-in, which enforces the use of regional forecasts for the designated control areas. Additionally, there is optional direct marketing of PV power based on forecasts for the respective PV power plants’ output. PV power is first offered on the day-ahead auction at the European Power Exchange. Subsequently, amendments based on updated forecasts can be made on the intraday market, when electricity might be traded until 45 minutes before delivery begins. Remaining deviations between scheduled and needed power are adjusted using balancing power. A similar procedure for California’s electricity market is described in Mathiesen, Kleissl, and Collier (2013). Also, Kleissl (2013) describes the stakeholder needs from the perspective of independent system operators and energy traders. Hence, accurate PV power forecasts at different spatial and temporal scales are very important for cost-efficient grid integration because large errors in the day-ahead forecast can cause either very high or negative prices on the intraday market and intraday forecast errors determine the need for costly balancing power