Evaluation of Weather Forecasting Uncertainties in Solar Thermal Power Plants Dispatch Optimization

Kurzfassung

The high penetration of renewable energy in electricity grids brings the role of balancing production and demand for Concentrated Solar Power (CSP) plants. Due to their thermal storage capability, CSP plants provide certain flexibility between collecting thermal energy from the sun and transforming it into electric energy by means of a power cycle. The dispatchability of CSP plants is based on the electricity demand and weather forecasts, to accurately schedule the electricity production over the following days. Such scheduling enables the participation of CSP plants in the balancing power market, ensuring the importance of such plants in the pathway to a highly renewable energy mix. Since weather forecasts include uncertainties, the dispatch schedule of CSP plants needs to take it into account. Modification of energy deliveries once scheduled is very limited and usually associated with drawbacks in form of penalties or reduced prices. Therefore, the uncertainty treatment in the energy delivery schedule is essential to ensure the optimization of solar thermal energy dispatch, with high dependency on type and accuracy of weather forecasts. Research on the precision of direct normal irradiance (DNI) forecast of different forecast methods was already performed, as shown by Law et al. (2014). Although, the impact of this accuracy on power scheduling and economic income of CSP plants is still a field to be studied. A dispatch optimization algorithm was developed at DLR, used to derive a CSP plant operation schedule for the upcoming 48 hours. It considers weather and electricity pricing forecasts with a special focus in the incorporation of uncertainty information. The strategy of the Dispatch Optimizer is based on a partitioned calculation between the optimization algorithm and the uncertainty processing. The optimization algorithm runs directly with all the possible weather scenarios, if the forecast is given as ensemble members, or deterministically with deviations, while the uncertainty is dealt as a post-processing. Therefore, the result of the optimization is a range of possible power schedules and the uncertainty post-processing is able to categorize single schedules suggestions according to the risk of meeting the promised energy delivery. This partitioned approach brings the possibility of including other parameters in the uncertainty treatment. The power schedule optimization considers not only the accuracy of the weather data but also electricity market characteristics. Therefore, several delivery strategies can be developed according to the decision maker’s point of view. This research shows as benefit the possibility of dealing with several types of weather forecast, such as probabilistic and deterministic methods, as well as enabling the evaluation of the quality of weather forecast data and its efficiency for optimizing the dispatch. Apart from that, it is possible to study the use of weather forecasts post-processing to improve accuracy of CSP plant operation schedule. Recent investigations on the economic benefit according to the type and accuracy of weather will be presented. It is expected that further results relating weather forecast accuracy and financial income expected from the CSP plant can be of great value for a better understanding on how to treat uncertainties for energy generation.