IT4Energy Stochastic Forecast
de: Das Diagramm zeigt die stochastische Vorhersage des Windenergieeintrags auf dem Dach unseres Instituts und vereint sowohl lokal gemessene Sensordaten als auch die Wettervorhersagen verschiedener Quellen.  en: The figure shows the stochastic forecast of available wind turbine energy production from the roof of our institute and combines locally measured sensor data with weather forecasts from other sources.  Fraunhofer FOKUS

Prognose als Dienstleistung

Prediction as a Service

Zur Unterstützung Ihrer Entscheidungen bieten wir Ihnen unseren Internet-Prognosedienst "smartPaaS" ("Smart Prediction as a Service") an. Dieser Dienst integriert unterschiedliche Datenquellen in unsere anwendungsspezifischen Prognoseverfahren: Sensoren, Bestandsdaten, kommerziell- und freiverfügbare Wetterdaten. Die resultierenden stochastischen Vorhersagen (Serien von Szenarien unterschiedlicher Wahrscheinlichkeit) von Energieflüssen, Preisen, Nachfragen, etc. stellen wir unseren Kunden und Partnern, vor dem Zugriff durch Dritte geschützt, zur Verfügung. Für den Fall, dass wir zur Erstellung unserer Prognosen auf Echtzeit-Sensordaten zugreifen sollen, bieten wir Ihnen Unterstützung bei der Wahl und geographischen Positionierung der Sensoren, um möglichst wirtschaftlich eine hohe Vorhersagequalität zu erreichen, gleichzeitig aber auch die Privatsphäre von Personen zu respektieren, getreu unserem 3S-Grundsatz ("Strategic Sampling und Sensing").

To support decision making in the industry we offer our internet-based forecasting service "smartPaaS" ("Smart Prediction as a Service"). This service integrates different data sources as inputs to our application-specific forecasting algorithms: M2M-enabled sensors (IoT data), facility and legacy data, commercial and open weather-related data, etc. The resulting stochastic (probabilistic) predictions consist of a series of future scenarios, each of which is assigned a given probability related to Energy flows, prices, demand, etc. Customized forecast are available to our partners via secure data pipelines. In the case of interest in our forecast and real-time sensor data platform, we can assist in the choice, deployment density and optimized positioning of on-site sensors, to achieve not only a better forecasting accuracy, but to save operational costs while guaranteeing quality of service and privacy according to our '3S' set of principles ('Strategic Sampling and Sensing').

IT4Energy Strategic Sampling and Sensing (3S) 
Strategic Sampling and Sensing (3S) Fraunhofer FOKUS

Strategic Sampling and Sensing (3S)

right: ubiquitous sensing at every resource usage/generation point (household, street, public areas) used for numerical forecasting 

left: 3S in which sensor placement is sparse, given privacy-aware limits of on sensor geographical density (shown by circle), leveraging spatial correlations in resource consumption to contain costs, and use targeted forecasts for multi-modal integration to achieve equivalent forecast accuracy

Dr. Florin Popescu

Dr. Florin Popescu received his Ph.D. in Mechanical Engineering from Northwestern University, USA in 2000, and is a Fraunhofer researcher since 2005. He has received distinguished awards such as the Marie Curie Excellence Team grant (2005-2009) for young researchers, in which he led a multi-disciplinary team of researchers in multimodal sensor-assisted robotics and led several other projects at Fraunhofer in renewables forecasting, machine learning approaches to industrial energy management and efficient deployment of smart sensors in urban environments. Among other research interests are automated and causal inference in dynamic systems. He is an author of over 30 peer-reviewed articles (in journals such as PlosONE, NeuroImage and proceedings such as ICAR and NIPS) and has co-edited a book on Causality in Time Series Research (JMLR W&CP series).