Common forecasting methods for photovoltaic power systems can be based a) on physical models and numerical weather predictions, b) data-driven, using PV monitoring data, c) estimating cloud motions from satellite images or sky cameras or d) hybrid-models using different aspects of above mentioned methods.
The basis of our forecast approach is a detailed physical model, allowing to consider very detailed technical characteristics of the PV systems, if this information is available. But it also allows for a simplified approach working with basic information on the PV systems, at slightly reduced accuracy. This model is driven by meteorological forecast data, originating from a numerical weather prediction service and delivers intra-day, up to 72 hours ahead forecasts.
Since our model follows a bottom-up approach, it models each PV system individually and reaches good accuracies for single-site forecasts (RMSE around 10%, considering day-time values only).
Background:In face of the growing number of variable renewable energy sources (particularly solar PV) in electrical networks, it becomes necessary to forecast this generation in a reliable and efficient way in order to support a better operation and planning of future energy distribution systems and to enable the integration of PV electricity into the short-term markets.
Benefits:Due to its detailed physical model and the bottom-up approach, LIST’s forecasting system is specifically powerful and accurate on single-site forecasts. Many data-driven forecasting algorithms, often applied to forecast the PV production of a portfolio of PV systems for utilities or other balance responsible parties, reach higher accuracies for large numbers of generators, distributed over a larger geographical area. While those forecast applications benefit from statistical effects (large numbers effect) and the smoothing of meteorological artefacts over larger regions, some applications require single-site forecasts for which the forecasting is much more challenging. Hence, our current forecasting approach provides state-of-the-art single-site forecasts for large portfolios of PV systems, which will in the future be complimented by a data-driven (machine learning based) module and the support of sky cam pictures for narrow geographical areas.
Applications:
LIST would be specifically interested in testing and implementing the developed approach in use cases of specific interest to single-site forecasts. Hence, implementing and testing our model in a precommercial demo in one of the following fields (or similar) would be favourable:
But also operation of virtual power plants or rather conventional forecast applications could be discussed.
IP Status:No patent, Copyright.
Desired business relationship
Technology development
Licensing
The Luxembourg Institute of Science and Technology (LIST) is a mission-driven Research and Technology Organisation (RTO) active in the fields of materials, environment and IT.
LIST develops competitive and market-oriented product/service prototypes for public and private stakeholders, and works across the entire innovation chain: fundamental/applied research, incubation, transfer of technologies.
By transforming scientific knowledge into technologies, smart data and tools, LIST empowers citizens in their choices, public authorities in their decisions and businesses in their strategies.
Create your free account to connect with Fabian Belin and thousands of other innovative organizations and professionals worldwide
Send a request for information
to Luxembourg Institute of Science and Technology (LIST)
Technology Offers on Innoget are directly posted
and managed by its members as well as evaluation of requests for information. Innoget is the trusted open innovation and science network aimed at directly connect industry needs with professionals online.
Need help requesting additional information or have questions regarding this Technology Offer?
Contact Innoget support
