This master's thesis deals with a design of backup power system, which will be using renewable energy sources, particularly solar energy. The first part describes the potential of solar power plant in the Czech Republic. The next parts describe types of photovoltaic systems, their components, design of photovoltaic systems and possibilities of controlling power consumption. Due to the fluctuating supply from renewable energy sources the thesis also deals with possibilities of predicting of the production electricity from these sources. In the practical part the design of backup power system for the house is done, which will also be working in summer as an optimizer for own consumption. A financial evaluation was done for this proposal.

Recent trends in technology scaling have shifted the processing paradigm to multicores. Depending on the characteristics of the cores, the multicores can be either symmetric or asymmetric. Prior research has shown that Asymmetric Multicore Processors (AMPs) outperform their symmetric (SMP) counterparts within a given resource and power budget. But, due to the heterogeneity in core-types and time-varying workload behavior, thread-to-core assignment is always a challenge in AMPs. As the computational requirements vary significantly across different applications and with time, there is a need to dynamically allocate appropriate computational resources on demand to suit the applications’ current needs, in order to maximize the performance and minimize the energy consumption. Performance/power of the applications could be further increased by dynamically adapting the voltage and frequency of the cores to better fit the changing characteristics of the workloads. Not only can a core be forced to a low power mode when its activity level is low, but the power saved by doing so could be opportunistically re-budgeted to the other cores to boost the overall system throughput. To this end, we propose a novel solution that seamlessly combines heterogeneity with a Dynamic Reconfiguration Framework (DRF). The proposed dynamic reconfiguration framework is equipped with Dynamic Resource Allocation (DRA) and Voltage/Frequency Adaptation (DVFA) capabilities to adapt the core resources and operating conditions at runtime to the changing demands of the applications. As a proof of concept, we illustrate our proposed approach using a dual-core AMP and demonstrate significant performance/power benefits over various baselines.

This work investigates how to improve wind power predictions using observational wind speed data. Measurements from ultrasonic anemometers (sonics) are available from five of the 22 wind energy turbines at the analysed wind farm in Beeskow, Germany (52°11’48'N, 14°13’E). In addition, measurements from a vertically pointing Doppler lidar (DL) at the Meteorological Observatory Lindenberg - Richard Aßmann Observatory located at a distance of 6 km from the wind farm are evaluated. The LoadManager® tool, developed by LEMSoftware, Leipzig, is used to perform wind power predictions based on different input data for forecasting horizons of 15 min and 30 min. Though wind power predictions have consistently improved in the last decade, persistent reasons for remaining uncertainties are sudden large changes in wind speed, so-called ramp events. The occurrence of ramp events at the wind farm has been investigated. Results on the seasonality of ramp events and their diurnal cycle are presented for multiple ramp detection thresholds. Ramps were found to be most frequent in March and April and least frequent in November and December. Furthermore, ramp events occur more frequently during the day than during the night and for most ramp detection thresholds up-ramp events are slightly more frequent than down-ramp events. For further analysis, the wind power prediction tool is fed with different wind velocity input data. The reference wind power predictions are based on data from a numerical weather prediction (NWP) model. Power predictions using observed wind speed data (DL, sonics) are compared to these reference predictions and evaluated according to two metrics: (i) the absolute difference between the observed and predicted power generation and (ii) the costs incurred due to necessary balancing services. It was found that, (i) the absolute power deviation can be reduced significantly compared to the reference by using power prediction setups based on sonic data. This improvement is even greater during ramp time steps. Power predictions based on the available DL data do not improve the absolute power deviation for the entire data set, albeit they do provide an improvement during down-ramp events. Considering (ii) incurred balancing costs, all power prediction setups based on observational data reduce the balancing costs compared to the reference. Sonic-based configurations yield 75-80% lower balancing costs than the reference and the DL-based setup results in 20% lower balancing costs. / Diese Arbeit untersucht, wie sich Windleistungsprognosen mit Hilfe von Windmesswerten verbessern lassen. Messungen von Ultraschallanemometern (sonics) an Gondeln von fünf der 22 Windenergieanlagen des untersuchten Windparks Beeskow, Deutschland (52°11’48'N, 14°13’E), sind verfügbar. Weiterhin sind Messungen des vertikalgerichteten Doppler Lidars (DL) am Meteorologischen Observatorium Lindenberg - Richard Aßmann Observatorium des DWD verfügbar, welches sich in einer Entfernung von 6km zum Windpark befindet. Das Programm LoadManager® der Leipziger Firma LEM-Software wird für Windleistungsprognosen mit verschiedenen Eingangsdaten für die Prognosezeiträume +15 min und +30min verwendet. Die Qualität von Windleistungsprognosen hat sich in den letzten zehn Jahren stetig verbessert. Unsicherheiten bleiben z.B. sogenannte Windrampen, schnelle, starke Änderungen der Windgeschwindigkeit. Das Auftreten von Windrampen am Windpark Beeskow wurde untersucht und die Ergebnisse werden für verschiedene Rampengrenzwerte vorgestellt. Am häufigsten treten Windrampen im März und April auf und am seltensten treten sie im November und Dezember auf. Außerdem wurden Windrampen häufiger tagsüber als nachts festgestellt. Für die meisten Rampengrenzwerte wurden etwas mehr Leistungsanstiege ('up-ramps') als Leistungsrückgänge ('down-ramps') gefunden. Für weitere Untersuchungen wurden Windleistungsprognosen mit verschiedenen Windgeschwindigkeitsdatensätzen durchgeführt. Als Referenz gelten Windleistungsprognosen auf Basis von Daten numerischer Wettervorhersagemodelle. Windleistungsprognosen auf Basis von Messwerten (sonics, DL) werden mit dem Referenzmodell verglichen und entsprechend zweier Metriken bewertet: (i) der absoluten Abweichung zwischen der vorhergesagten und beobachteten Stromerzeugung und (ii) der für Abweichungen anfallenden Regelenergiekosten. Die Ergebnisse zeigen, dass (i) die absolute Abweichung verglichen mit der Referenz signifikant reduziert werden kann, in dem man Messwerte von sonics für die Leistungsprognose verwendet. Dabei ist die Verbesserung während Windrampen größer als für den gesamten Datensatz. Windleistungsprognosen auf Basis von DL-Daten zeigen keine Verbesserung der Abweichungen für den gesamten Datensatz, jedoch eine signifikante Verbesserung während Leistungsrückgängen. Betrachtet man (ii) die anfallenden Regelenergiekosten, resultieren alle auf Messwerten basierenden Leistungsprognosen in einer Reduktion der Kosten verglichen mit dem Referenzmodell. Windleistungsprognosen auf Basis der Gondelmessungen reduzieren die Regelenergiekosten um 75-80% und Windleistungsprognosen auf DL-Basis ergeben im Mittel etwa 20% niedrigere Regelenergiekosten.

The wind power prediction plays a vital role in a wind power project both during the planning and operational phase of a project. A time series based wind power prediction model is introduced and the simulations are run for different case studies. The prediction model works based on the input from 1) nearby representative wind measuring station 2) Global average wind speed value from Meteorological Institute Uppsala University mesoscale model (MIUU) 3) Power curve of the wind turbine. The measured wind data is normalized to minimize the variation in the wind speed and multiplied with the MIUU to get a distributed wind speed. The distributed wind speed is then used to interpolate the wind power with the help of the power curve of the wind turbine. The interpolated wind power is then compared with the Actual Production Data (APD) to validate the prediction model. The simulation results show that the model works fairly predicting the Annual Energy Production (AEP) on monthly averages for all sites but the model could not follow the APD trend on all cases. The sensitivity analysis shows that the variation in production does not depend on ’the variation in roughness class’ nor ’the difference in distance between the measuring station and the wind farm’. The thesis has been concluded from the results that the model works fairly predicting the AEP for all cases within the variation bounds. The accuracy of the model has been validated only for monthly averages since the APD was available only on monthly averages. But the accuracy could be increased based on future work, to assess the Power law exponent (a) parameter for different terrain and validate the model for different time scales provided if the APD is available on different time scales.

No description available.