Suchergebnisse

A strong increase of onshore and offshore wind power capacities is an official political target in Germany and other countries. The wind energy shares therefore rise in many power systems. Wind power generation has other characteristics than the power generation by conventional power plants. The wind is a natural resource that is fluctuating. The meteorological dependency leads to a limited predictability of the available power. A third aspect is the concentration of wind farms at locations with high wind yields as in the North of Germany. From a methodological point of view, the thesis focuses on the analysis of the three aspects with regard to the power system operation and the development of related modelling approaches. This especially refers to the application of a stochastic optimization model for the system analysis and to the simulation of wind power generation and wind power forecasts. The application orientated focus is on a scenario analysis of the German power system in 2020. The analysis aims at the identification of promising system adaptations that lead to an improved wind power integration and a more efficient power system operation. Before the model presentation, the importance of the three aspects above is discussed giving the basics for the latter modelling. It is shown that the residual load fluctuations are increased by the wind power generation, especially if they are related to the residual load levels. The flexibility of thermal power plants is also regarded here. An analysis of operational uncertainties shows the importance of wind power forecast errors in relation to load forecast errors. The DC load flow model and characteristics of the transmission grid are explained. A stochastic market model is presented that allows an integrative analysis of the wind power integration. One characteristic of the optimization model is the application of a rolling planning so that forecast errors can be specifically considered. A main modification of the model compared to earlier model versions is given by the representation of grid constraints. A grid reduction approach is developed that reduces the transmission grid to a simplified structure that is applied in the market model. The grid reduction approach is based on a comparison of DC load flow solutions in the reduced and unreduced grid. Additionally, an approach for the calculation of tertiary reserves is given. The approach considers the wind forecast quality and combines probabilistic elements with an optimization. The simulation of wind power generation and forecasts combines different analyses and methods. General quantitative relations between the variability of wind power generation and the geographical region size are derived. The equations are applied in the simulation of wind power generation that is based on adapted wind power curves. The adapted power curves consider regional smoothing effects in the transformation of wind speed to wind power. The simulation results reflect the high variability of the concentrated offshore wind power. For the simulation of the wind power forecasts, a scenario generation method based on moment matching is presented that allows simulating non Gaussian distributed forecast errors and their correlations. The results of a statistical analysis of measured forecast errors are used in the simulation. An empirical relation between error correlation and geographical distance is for example given. The German forecast quality that is simulated for 2020 assuming an improvement of forecasting by 20% is, related to the installed capacity, similar to the one of today due to the high spatial concentration of the offshore capacities. For the scenario analysis of the power system in 2020, the power plant portfolios of twelve German regions and other parameters are derived based on different sources. This includes reserve requirement values and reduced grid parameters that are calculated by the methods mentioned above. The results show that, in the regarded scenario, 3% of the yearly wind energy cannot be integrated into the system. They are curtailed nearly exclusively due to transmission constraints. The network congestions also lead to high differences between the regional electricity prices. The yearly costs of wind forecast errors amount to circa 180 million Euros or 1% of the operational system costs. The model results thereby indicate a large cost saving potential by risk management methods. Based on scenario modifications, integration measures related to CAES capacities, demand side management and more flexible power plants as well as infrastructural changes by grid expansions and an adapted geographical allocation of power plants are analysed. The importance of a stochastic modelling approach for the evaluation of flexibility related scenarios is shown. The comparison of the integration measures identifies infrastructural changes as most efficient system improvements whereas the benefits of CAES capacities are small. Assuming a grid without any transmission constraints, the yearly system costs are reduced by one billion Euros. A limited grid upgrade leads to 10% of this cost reduction. Similar cost savings are achieved by adapting the geographical locations of the power plants. Adjusting the generation to the grid is therefore a promising alternative to grid expansions especially considering the long processes that are involved with new transmission lines. A market design with regional electricity prices would give related incentives. ; Ein starker Ausbau der Windenergie, onshore und offshore, ist ein erklärtes politisches Ziel in Deutschland und anderen Ländern. Der Windenergieanteil nimmt folglich in immer mehr Stromsystemen zu. Die Stromerzeugung durch Windenergie weist andere Eigenschaften auf als sie beim Betrieb konventioneller Kraftwerke gegeben sind. Die Erzeugung beruht auf der natürlichen Ressource Wind und ist daher fluktuierend. Die meteorologische Abhängigkeit führt auch zu einer begrenzten Prognostizierbarkeit des verfügbaren Stromes. Ein dritter Aspekt ist die Konzentration der Windenergieanlagen an windreichen Standorten wie im Norden Deutschlands. Die Arbeit verfolgt zunächst das methodische Ziel, diese drei Aspekte bei der Analyse des Strombetriebs greifbar zu machen und geeignete Ansätze für die Strommarktmodellierung zu entwickeln. Dies betrifft insbesondere die Simulation der Windstromerzeugung und Windstromprognosen und die Anwendung eines stochastischen Optimierungsmodells zur Systemanalyse. Außerdem wird das anwendungsorientierte Ziel verfolgt, verschiedene Szenarien des deutschen Stromsystems für das Jahr 2020 zu untersuchen. Vielversprechende Systemanpassungen für eine verbesserte Integration der Windenergie und einen effizienteren Strombetrieb sind dabei zu identifizieren. Vor der Methodenentwicklung und Anwendung des Strommarktmodells wird zunächst die Bedeutung der drei obengenannten Aspekte erörtert und die Grundlage für dianpassungen hinsichtlich des Einsatzes von CAES Speicherkraftwerken, Demand Side Mangagement und flexiblerer Kraftwerke sowie Änderungen der Infrastruktur durch Netzausbauten und einer angepassten geographischen Allokation von Kraftwerken untersucht. Die Ergebnisse belegen die Vorteile eines stochastischen Marktmodellierungsansatzes für die Evaluierung flexibilitätsbezogener Integrationsmaßnahmen. Der Vergleich der Integrationsmaßnahmen identifiziert infrastrukturelle Maßnahmen als wirksamste Systemverbesserung, während der Nutzen zusätzlicher CAES Speicherkapazitäten gering ist. Unter der Annahme eines Netzes ohne Übertragungsengpässe reduzieren sich die jährlichen Systembetriebskosten um eine Milliarde Euro. In einem moderateren Netzausbauszenario lassen sich nur 10% dieser Einsparung realisieren. Eine Kostenreduktion in ähnlicher Höhe wird durch eine angepasste Standortwahl neuer Kraftwerke erreicht. Eine Anpassung der Stromerzeugung an die Netzsituation ist demnach eine vielversprechende Alternative zu Netzausbauten, die vor allem auch wegen der langwierigen Prozesse beim Bau neuer Netzleitungen interessant ist. Ein regionales Preismodell würde dafür Anreize schaffen.