· · · Institutional Repository

Voorstudie en uitwerking van een centrale van ongeveer 350 kW nabij Hoge Bat, keuze uit diverse locaties, vergelijking diverse principeoplossingen. Bepaling maatgevende belastingen, dimensionering en economische berekening

The history books of tomorrow will remember the start of this millennium by the global energy dilemma. The literature will portray the initial skepticism towards the first signs of global warming and the depletion of unsustainable energy sources. It will lay out how the evolution of mild symptoms into hard realities -such as the exponential increase in droughts, hurricanes and other natural disasters-drove these polarized opinions to a unified global belief that radical measures in energy policy needed to be taken. The accounts will continue by describing how the global discussion then moved to the implementation of sustainable energy alternatives. And they might explain how this drove the competition between solar and wind energy industries, which in turn pushed the development of their technologies. Even though dynamic substructuring might not star in those passages, I believe it will contribute to the technological evolution needed to sustainably meet tomorrow’s energy demand. This includes the field of wind turbines. Dynamic substructuring is in essence a linear structural dynamic analysis methodology that models a structure by the assembly of a set of substructure models, hence segmenting the normal linear model. By the implementation of this approach, components can be validated individually and reduction techniques can be tailored to the characteristics of and interaction between the components. Moreover, processing costs can be reduced via parallel computing and the ability to interchange component models without having to re-build the entire dynamic structural model. However, when operational analyses such as time integration are required and the model or its components are expected to undergo large rotations or displacements, the dynamic substructuring methodology needs to be integrated into a formulation that includes rotational effects, such as the floating frame of reference formulation (FFR). This formulation places the linear structure model into a floating reference frame. The body’s deformations within this frame remain linear, but the frame’s kinematic description is highly non-linear. This often leads to major reductions in processing efficiency, diminishing the competitive advantage that dynamic substructuring offers in comparison to other formulations. In the case of the operational analyses of multi-megawatt wind turbines on the other hand, rotation forces on the important turbine sub-systems tend to be relatively low. Hence the engineer’s gut feel leads to the belief that the floating frame of reference formulation can be significantly simplified. By (partly) linearizing the terms in this formulation, processing costs might significantly reduce at the loss of only little accuracy in the analyses. This leads to the main research question of this thesis: To what extent can the rotational effects be simplified for various wind turbine operational analyses without significantly impacting their dynamic characteristics? In order to answer this question, a simple wind turbine dynamic model is constructed in the floating frame of reference formulation. The terms in this formulation are subsequently simplified by various approaches and evaluated in terms of accuracy and processing costs in three different time simulations. This findings of the analysis will indicate which simplification approach will provide accurate results at minimal processing cost. The results indicate that the processing cost of the 2.3 megawatt turbine model can be reduced to less than one percent with respect to the reference model in steady operation, at negligible accuracy loss. In more transient cases, such as the simulation of startup or (emergency) shut down, the degree of simplification needs to be reduced to maintain the accuracy of the model. this leads to a smaller reduction in processing cost, now to twenty percent in of the original processing costs. The research demonstrates that the simplification of rotational effects in turbine models is feasible and has a significant effect on the processing times of the codes. The thesis hence recommends further development of the simplified FFR formulations. The formulations could be implemented into dynamic substructuring tools, allowing the analysis of operational dynamic interaction of a system’s subcomponents at competitive processing costs. Also, turbines of other sizes could be investigated, possibly leading to a presentation of the simplifications in non-dimensional form. In result, tailored simplifications of rotational effects can be applied to any wind turbine model. From the theoretical perspective the thesis recommends further research into simplifying the FFR formulations using other parametrization techniques, such as Euler parameters and Rodriguez parameters. These might pose new opportunities in terms of processing efficiency. The optimization of dynamic turbine modeling will contribute to speedier design iterations that result in more cost effective turbine designs. With a little bit of luck this industry might consequently turn wind into a leading global supplier of energy. And who knows, the odd history book might mention the contribution of dynamic substructuring to this process.

Deze deelstudie is een gedeelte van de totale afstudeeropdracht, die het volgende inhoudt; "het onderzoeken van het OPAC-plan (Ondergrondse Pomp Accuulatie Centrale) met een grootte 1200 MW, welke gebruikt zal worden voor opslag van windenergie alswel voor piek-shaving". In deze deelstudie wordt een deelonderzoek gedaan naar het (totaal)procesrendement van het OPAC-systeem dat voor een groot gedeelte de effectiviteit van dit systeem bepaalt. Er wordt aangenomen, dat het OPAC-systeem alle pieken van de vraag- en aanbodcurve zo goed mopgelijk moet kunnen volgen een e.a. met een zo hoog mogelijke efficiency (rendement). D.m.v. verschillende vermogensgrootte van de (Francis) pomp-turbine eenheden zal worden getracht om de efficiency (totaalprocesrendament) van het OPAC-systeem te verhogen t.o.v. de huidige stand van onderzoek. Daarbij wordt in deze deelstudie alleen gekeken naar de normale bedrijfsvoering, die onder optimale omstandigheden gebeurt, en er wordt geen rekening gehouden met bijv. omschakelverliezen en aanloopverliezen. Dit zal in de deelstudie van W.A. de Haan nader bekeken worden.

The process of designing a wind turbine blade involves many disciplines and is of an iterative character, making it cumbersome and sometimes inefficient. In this thesis an increased understanding of the physical and modeling considerations required for a better aero-elastic design of a blade is sought. To this end, an aero-structural model is developed and placed in an optimization procedure, where the mechanisms driving the optimizer can be analyzed. The model consists of a Blade Element Momentum analysis of the flow; and a process that dimensions the internal structure of a blade for a set of aerodynamic loads and design requirements. The optimizations use Genetic Algorithms and are done in a parametric way, increasing the design space and changing the objective function step by step. The studied parameters are the chord and twist distributions and the airfoil thickness of the outboard half of the blade. The optimizations are carried out with the objective of improving the design of the DOWEC reference turbine for power performance, blade mass and a combination of the two. The results of the optimizations show the mechanisms that must be taken into account when doing the preliminary aero-structural design of a wind turbine blade. In particular, one of the results indicates that the mechanisms for an enhanced power performance design and reduced blade mass are not always counteracting and can actually, with a good understanding of the physical phenomena, be used for an improved design in both aspects.

This report describes the graduation project of Hengfeng Chi, a master student of Inte- grated Product Design at the faculty of Industrial Design Engineering TU Delft. The project is to design an urban wind turbine with appreciated appearance in the context of coastal area in Scheveningen, in collaboration with Actiflow BV. The project consists of four phases: Technology and product semantic analysis, Context research, Product concept development, Concept finalization. In the end, an innovative urban wind turbine was designed, which would be further engineered and probably to be built on the Pier of Scheveningen in a few years.

Direct drive generators applied to large wind turbines present some problems, such as very heavy and expensive price. The use of magnetic bearings has a possibility to reduce the weight of the direct drive generator. The control system for such magnetic bearings is considered. In the beginning, the thesis discusses the problems of direct drive generators in large wind turbines, introduces a hybrid concept of active magnetic bearings, gives a demonstrator of magnetic bearings used in this project, and presents a basic control system of active magnetic bearings. For the purpose of support such magnetic bearings in wind turbines, this thesis gives a complete control system. This control system includes electrical circuits and decentralized control method. The implementation of the electrical circuits is distributed into two PCBs. The decentralized control method is designed with six PID controllers. Finally, in order to improve the stability of the system, the H-infinity control method is suggested to magnetic bearing system in wind turbine applications.

In dit hoofdontwerp is gepoogd om uitgaande van de funktionele eisen en de randvoorwaarden te komen tot het ontwerp van de turbine-eenheid. Allereerst wordt de getijdencentrale gesplitst in een aantal komponenten, elk met een specifieke funktie, en wordt het trace bepaald. Vervolgens wordt de komponent met de genererénde funktie nader uitgewerkt. Nadat is gekozen voor prefabrikage en een aantal principeoplossingen zijn geformuleerd, wordt aan de turbine-eenheid een konkretere vorm gegeven, hetgeen resulteert in een beperkt aantal alternatieven. De keuze hieruit wordt gemaakt door het beschouwen van enkele grenstoestanden. Tenslotte worden van het gekozen alternatief een uitgebreider aantal grenstoestanden, zowel in bouw- als gebruiksstadium bekeken, evenals een beknopte foutenboom.

Voor de opwekking van duurzame energie neemt het aandeel offshore windturbines stelselmatig toe, dit door de vele voordelen ten opzichte van landelijke windturbines. Het principe is overal hetzelfde; in de gondel wordt de mechanische energie wordt omgezet in elektrische energie. Hoe hoger de mast, hoe hoger de energieproductie omdat de windsnelheid in hoogte toeneemt. De uitvoering van de mastconstructie boven het wateroppervlak is vrijwel gelijk voor alle offshore windturbines maar onder de waterlijn kan de fundering verschillen van plaatst tot plaats. Dit naargelang de diepte, draagkracht zeebodem, optredende belastingen veroorzaakt door drijvend ijs golven, waterstroming. De fundering van offshore windturbines in de Noordzee wordt meestal uitgevoerd in de vorm van een monopaal. Waar ter plaatse van de bodem een bodembescherming wordt toegepast. In dit verslag wordt aan de hand van vereenvoudigde rekenmethoden gekeken of het weglaten van een bodembescherming rondom de paal grote gevolgen heeft op de dynamische response van de constructie als gevolg van erosie. Er wordt hierbij gekeken hoe groot de verandering is van de eigen frequentie nabij of in het excitatieveld, omdat dit een goede indicatie geeft betreffende de mogelijke gevolgen hiervan. Er zijn een viertal belangrijke dynamische excitatiefrequenties die van invloed zijn op de dynamische response van de constructie. De uitgevoerde excitatie door wind rondom de mast is in dit verslag buiten beschouwing gelaten. Indien de eigenfrequentie dichter in de buurt komt van een excitatiefrequentie dan heeft dit invloed op de levensduur omdat bij gelijkblijvende dimensionering de kans op vermoeiingschade groter wordt. Dit kan extra onderhoudskosten met zich meebrengen of een beperking in levensduur betekenen. Naast het effect van lokale erosie is er gekeken naar het dynamisch effect bij een verschil in funderingsgrond zoals zand of klei en anderzijds het gebruik van een materiaal anders dan staal, waarbij gekozen is voor beton. Voor de gehele studie wordt gebruik gemaakt van een monopaal waarvan de dimensies grofweg zijn bepaald aan de hand van de belangrijkste belastingen zoals het eigen gewicht van de paal, gewicht gondel en rotor alsook belastingen als gevolg van water en luchtstroming. Diepgaande detaillering is hier niet van toepassing omdat het hierbij gaat om een vereenvoudigd model. Omdat er wordt gekeken naar het materiaal beton, wordt er in deze studie uitgezocht hoe een dergelijke betonnen constructie gefundeerd en opgebouwd kan worden. Daarbij kwam naar voren dat een zogenaamde "boor variant" in de eerste ontwerpcyclus als beste naar voren kwam. De betonnen paal van 70 meter hoog (t.o.v. waterspiegel) en een diameter van 8.9m die een waterdiepte overbrugt van 20m is opgedeeld in verschillende losse ringvormige elementen die door middel van 47 staven wordt voorgespannen. Om de elementen in het funderingsgedeelte voldoende de grond in te krijgen, worden ze een voor een opeengestapeld waarbij de hoeveelheid grond in het binnenste gedeelte wordt losgemaakt door een combinatie van spoelen onder hoge druk, boren en afzuigen. Op deze manier kan men de puntweerstand vrijwel elimineren en kan men de wrijvingsweerstand overwinnen door enkel gebruik te maken van het eigen gewicht. Conclusies In de studie wordt aangetoond dat de invloed van een erosiekuil beperkt van aard is op de eigen frequentie, daarbij moet nog worden opgemerkt dat de eigen frequentie bij het genomen voorbeeld nog steeds in de excitatieband bevindt van de golfbeweging en de rotorfrequentie. De situatie blijft dus voor zover "ongunstig". Om echt te weten te komen of de situatie werkelijk ongunstig is gebleven zou men een meer gedetailleerde berekening moeten maken waarin wordt gekeken naar de kansverdeling waarbij een bepaalde excitatie optreed.

To meet the growing demand for sustainable energy, existing wind turbines have to be made more powerful and efficient. One of the research topics to achieve this, is the aero-elastic stability limit of the blades. This can be analyzed by the vibrations that occur on the blades during normal operational conditions. In this study high speed photogrammetry is used to measure blade vibrations of a model wind turbine. By placing targets on the blades of the wind turbine and taking images with two synchronized high speed cameras, the motion of the blades is captured. High speed cameras produce a large number of images, making manual target measurements impractical. In this study, it is investigated how target measurements in high speed images can be automated and what accuracy can be obtained for the reconstructed object coordinates. To automate the target measurements, a method is developed based on target detection and tracking. The targets in the images are detected using a threshold histogram segmentation. Since the targets describe a circular motion a circle is used as model to track its position in the sequence of frames. To measure the targets with subpixel precision, the edges of the targets are detected and a circle is fitted to the edges. To reconstruct the object coordinates a Direct Linear Transformation (DLT) is used. The DLT equations are linear and can easily be solved with standard Least Squares Estimation. The developed methods were used for an experiment in a wind tunnel, whereby images of a model wind turbine were made with an acquisition frequency of 500 Hz and a rotational speed of 260 rpm. Retro-reflective targets were put on the blades to identify the measuring points. Using the developed methods, the targets were successfully measured and tracked in the images of the experiment. Using the targets measured in the images, 3D object coordinates were calculated with an accuracy of 1.32 millimeter. The vibrations of a target were obtained by means of Principal Component Analyses (PCA). Using PCA, the track of coordinates of a target was transformed to a new coordinate system. The xy plane of this new coordinate system coincided with the rotation plane of the target and the z axis is perpendicular to it, containing the vibrations. The measured vibration had a maximum amplitude of 6 mm, with an accuracy of 0.56 mm in the vibration component. From the performed experiment with the model wind turbine, it can be concluded that the blade vibrations can be measured using high speed photogrammetry.

Wind energy promises to become an important source of energy in the near future. Penetrating large-scale wind power into power systems presents a lot of challenges to power system operators, generation companies and wind turbine manufactures. In order to design less expensive and more efficient wind turbines, manufacturers have tried a lot of possibilities. However reliability is also an important issue. For this reason much attention needs to be paid to the reliability improvement. In the beginning, this thesis introduces and discusses modern wind turbines, gives failure rates to express the reliability, analyzes problems and the most critical subassemblies of wind turbines, and gives the possible methods to make wind turbines more reliable. For electrical part of wind turbine, generator is a crucial component. In this thesis a fault tolerant energy conversion system was designed to meet the specification. With certain faults the system can continue operating and output power. Therefore the reliability of wind turbines is increased. Finally the design procedure, and comparisons of this fault tolerant generator system and conventional generator system are given.

The trend of highly energy-efficient, low carbon economies leads to high targets for renewable energy production in 2020, with main contributions by onshore wind, offshore wind and biomass. Many offshore wind farms have to be built to meet these targets. These wind turbines will be built farther and farther away from the coast. When a failure occurs, the maintenance crew has to travel farther and therefore needs a longer good weather window. Since these longer windows occur less often than small good weather windows, waiting times until a good weather window appears will probably be larger for these wind turbines. As long as the maintenance crew can not reach the failed turbine, this turbine cannot produce power and will thus not provide revenue. To reduce the downtime caused by these weather waiting times, a service island can be used. This is a fixed point at sea from where maintenance can take place, where engineers can stay and spare parts can be kept. To determine an optimal location for such a service island, the problem can be written as the Weber problem. The problem is to find the point for which the weighted sum of distances to given points is minimized. Using the wind farms or wind turbines as given points, and the ratio of weighted power as weights, this problem can be solved with the Weiszfeld algorithm. This algorithm finds a solution by using a converging sequence, based on the first-order necessary conditions for a point to be optimal. Since the objective function is strictly convex, the Weiszfeld algorithm finds the global minimum. In this solution all the maintenance for all wind turbines is done from this location, even of the ones that are closer to a port than to the service island. Because transportation time is to be minimized this might not be the best location for the purpose of a service island, and for this reason a generalization of the problem is considered. By adding more service locations, a port can be incorporated in the solution and each wind farm or wind turbine is assigned to one service location. The problem is to find the points for which the sum of distances to the given points, with the corresponding allocation, is minimized and is known as the unconstrained continuous location-allocation problem. The objective function of this problem is neither convex nor concave, which can cause a large number of local minima. If the number of new locations is unknown, the problem is NP-hard, but if the number of new locations is known, the problem is polynomial solvable. The problem can then be solved with the MALA algorithm. This algorithm allocates each wind farm to the closest service location and then solves the problem for each service location individually with the Weiszfeld algorithm. This procedure is repeated until no further reduction in total cost can be made. With the solutions of the MALA algorithm, the maintenance during the lifetime of the wind turbines is simulated. Hereto failure rates en repair time distributions are used to simulate failures and repairs. Four maintenance categories are considered for failures of different components of the wind turbine. It is assumed that there is only one maintenance crew available and that each maintenance category has its own vessel with its own weather specifications. Wind and wave heights of the period 2001 until 2010 are used to simulate the wind and wave pattern. This simulation gives insight into the availability of wind turbines and the influence of service islands. The presence of a service island increases the availability of the wind turbines and increases the number of repairs. A cost comparison is made to determine when a service island is profitable, considering the cost of an island, the kilowatt hour price and the losses caused by the downtime of turbines. The maximum investment budget increases when the kilowatt hour prices increases, but also the total cost involved increase. Therefore the profitability of a service islands depends on the (expected) kilowatt hour price.

Optimal design of support structure including foundation and turbine tower is among the most critical challenges for offshore wind turbine. With development of offshore wind industry heading for deeper ocean areas, new support structural concept such as a full lattice tower could be proven to be more advantageous than others when consideration of cost, safety and even environment aspects, etc are taken into. This thesis first gave introduction of present industrial applications of hybrid support structural concept combing a lattice foundation and monotower in relatively deep water areas before the presenting and introduction of challenges with the transition piece component. Conceptual model of transition piece design for a full lattice tower support structure proposal was discussed extensively which included consideration of structural form, functional requirement, mechanical condition, etc. A mechanical model of transition piece with regards to boundary condition and load conditions was also provided. Design and analysis of two different types of transition piece models under various load conditions were performed during preliminary design and with conclusion drawn, a refined final design of transition piece model for the full lattice tower support structural concept which has also included more practical aspects was assessed through investigation of its performance under varying load conditions and different load cases. This refined final design was found to be the most optimal design fulfilling all relevant requirements at a comparable structural cost. Conclusion and recommendation was therefore given in the last part. This thesis work is serving for a novel proposal of support structural concept for future offshore wind industry and relevant present experience is in fact nonexistent. The work applied present industrial information of hybrid support structure as basis along with consideration of offshore wind turbine structural and operational mechanism and their respective requirement. Structural analysis of transition piece design was conducted by means of finite element analysis technique and structural load conditions were simulated through up-to-date numerical modeling code for offshore wind turbine structure. Due to absence of relevant verification sources, advice and correction of the thesis content is much appreciated.

The trend in wind energy is towards large offshore wind farms. This trend has led to the demand for high reliability and large single unit wind turbines. Different energy conversion topologies such as multiple stage geared generators, single stage geared generators and gearless (direct drive) generators have been coupled to the wind turbine in the last decades. Direct drive generators based on permanent magnet technology have a high efficiency, less components and lower speeds that can translate to high energy yield and less maintenance demand. However, the mass of such direct drive generators can be significantly higher than of other geared generators. This thesis focuses on the issue of mass reduction of large direct drive generators in wind turbines. The main objective of this thesis is to investigate methods of mass reduction in large direct drive generators in wind turbines especially mass reduction by introducing structural flexibility in the generator. This allowance in structural flexibility is enabled by the use of magnetic bearing technology.

Issues related to environmental concern and fossil fuel exhaustion has made wind energy the most widely accepted renewable energy resource. However, there are still several challenges to be solved such as the integrated design of wind turbines, aeroelastic response and stability prediction, grid integration, offshore resource assessment and scaling related problems. While analyzing the market of wind turbines to find the direction of the future developments, one can see a continuous upscaling of wind turbines. Upscaling is performed to harness a larger resource and benefit from economy of scale. This will pose several fundamental implications that have to be identified and tackled in advance. This research focuses on investigating the technical and economical feasibility and limits of large scale offshore wind turbines using the current dominant concept, i.e. a three-bladed, upwind, variable speed, pitch regulated wind turbine installed on a monopile in an offshore wind farm. Thus, the objective of this research is to investigate how upscaling influences the offshore wind turbines. Specifically, following questions are of interest: 1. How do the technical characteristics of the larger scales change with size and can these technical characteristics appear as a barrier? 2. How does the economy of the future offshore wind turbines change with size? 3. What are the considerations and required changes for future offshore wind turbines? To address these questions, a more sophisticated method than the classical upscaling method should be employed. This method should provide the detailed technical and economical data at larger scales and address all the design drivers of such big machines to identify the associated problems. However, interdisciplinary interactions among structure, aerodynamics and control subject to constraints on fatigue, stresses, deflections and frequencies as well as considerations on aeroelastic instability make the development of such a method a cumbersome and complex task. Among many different methods, integrated aeroservoelastic design optimization is found to be the best approach. Therefore, the scaling study of this research is formulated as an multidisciplinary design optimization problem. This method enables the design of the future offshore wind turbines at the required level of details that is needed to investigate the effect of size on technical and economical characteristics at larger scales. Using this method, 5, 10 and 20 MW wind turbines are designed and optimized, including the most relevant design constraints and levelized cost of energy as the objective function. In addition to the design of these wind turbines, the method itself shows a clear way forward for the future offshore wind turbine design methodology development. Based on these optimized wind turbines, scaling trends are constructed to investigate the behavior of a wind turbine as it scales with size. These trends are formulated as a function of rotor diameter to properly reflect the scale. Loading, mass, cost and some other useful trends are extracted to investigate the scaling phenomenon. Blades and tower as the most flexible load carrying components are examined with more attention. Using these results, the challenges of very large scale offshore wind turbines up to 20 MW range are explored and identified. These results demonstrate that a 20 MW design is technically feasible though economically not attractive. Therefore, upscaling of the current wind turbine configurations seems to be an inappropriate approach for larger offshore wind turbines.

This study investigates the possibility to build and operate a Tidal Power Plant (TPP) in Saemangeum in South Korea. The objective of the study is to investigate if a TPP would be technically possible and economically feasible for Saemangeum. This has resulted in a design for a Tidal Power Plant. All decisions were made taking into account possible future shifts in facts in economy, energy prices and spatial planning ideas. The most important future functions will be polders for agricultural, industrial or residential purposes and a fresh water basin. A design without one of these functions is most likely to be rejected during the decision process. Existing plants and their performances have been studied as well as several feasibility studies made throughout the years, including the Sihwa project, a single, low basin plant which is at present under construction in South-Korea. A single, high basin and a single, low basin scheme are the most interesting schemes for the Saemangeum case. A bulb turbine is the most suitable type of turbine for tidal power. A Storage Area Approach Model was made to predict the energy output. The possibility to vary the basin area (including the depth-storage relationship) is built in, as well as the possibility to vary the number, efficiency and rated head of turbines (related to the head and flow velocities) and additional sluicing capacity. Also the construction costs, a Net Present Value calculation and the annual energy output are some of the output parameters of the model (with varying discount rate, future energy price). For various reasons it was decided that a layout with the basin area of 114 km2 at MSL must be selected and equipped with a low basin plant: This layout contains a large inter tidal zone, both a fresh water basin and polders, and it turns out to be the economically most feasible option. The selected layout contains a fresh water basin of 90 km2. Generation will only take place in one direction, from sea to basin (flood generation). No extra sluicing capacity is needed, as the 300 meters of present sluice length are sufficient. The powerhouse will be equipped with 18 bulb turbines (runner diameter 7.5 m) and generators, having a total installed capacity of 142 MW. The plant will generate approximately 454 GWh per year and the construction costs will amount to 286 Million US$. For the construction process a cofferdam consisting of circular cells and connecting cells is built around the future TPP power house location. At the barrage side the circular cells have a diameter of 31.8 m and at the basin side the circular cells have a diameter of 25.2 m. The given parameters have been determined by an economic optimisation. The numbers given above are based on assumptions: the plants design has a maximal Net Present Value after 40 years, with a discount rate of 4 %, an annual rise of energy price of 4 % and an actual energy price of US$ 0,03 per kWh. This production price has been set at US$ 0,03 per kWh, because the plant is expected to be able to compete with other energy sources. Under these assumptions tidal power in Saemangeum turns out to be attractive and feasible. The Net Present Value after 40 years of operation is US$ 187.6 Million for a discount rate of 4 % and an expected annual rise of energy price of 4 %. The break even point will be reached after 25 years of operation and the Internal Rate of Return is 6.5 %.

In the need for more green energy a prominent role is reserved for wind energy. Offshore wind energy in deeper waters capitalises on more efficient wind properties and increased public acceptance compared to onshore wind energy and wind farms close to shore. In the coming years the offshore wind market is expected to evolve rapidly, especially in the deeper water range of thirty to sixty meter. In a business case preceding to this study as first reference a jacket type substructure was designed for a 6 MW turbine in a water depth of sixty meter. The goal of this thesis is to reduce the cost of this reference design in total use of material and assembly. Also the transportation and installation of the substructure are taken in consideration. First a step back is taken to reconsider the structural concept of the reference design. Several substructure concepts, like tripods and straight-leg jackets, have passed the review and firstly qualitative weighed against primary criteria in a Multi Criteria Analysis and subsequently by FEM based in-place analysis. The outcome of the total substructure weight and natural frequency with respect to frequency of wave loading and turbine excitement determined the decision to further investigate a three-leg and four-leg battered jacket. Thereto a fatigue analysis was performed. The calculation method used at the original reference design to determine the total fatigue damage due to turbine and wave loading was proven to be too optimistic and therefore modified. In relation to the reference design several optimisations have been proposed, including applying a horizontal brace just above mudline level, applying double sided butt welds and adopting K-bracing instead of X-bracing. Here the four enclosed pictures can be placed. (number 1 upmost left, number 4 upmost right) Finally, four designs have been worked out; the reference design (without optimisations), an optimised four-leg jacket, four-leg jacket with k-braces and a three-leg jacket. The total assembling cost of each design is calculated by considering the handling time and the welding volume with corresponding welding time of each weld. Together with the material use the total fabrication cost is assessed. The jacket shall be transported offshore by a standard North Sea barge. The dimensions of this barge potentially enable the transportation of three four-leg jackets and four three-leg jackets. Depending on the wind farm location this may lead to reduction of one tug and transport barge case of the three-leg jacket. Further consequence of the three-leg jacket is that a foundation pile less needs to be driven. Thereto is the installation time of the three-leg jacket reduced, resulting in less installation cost. By combining fabrication, transport and installation cost it is possible to compute an overview for substructures cost in a complete wind farm. Final conclusion is that the fabrication cost are decisive compared to the installation and transport cost. The four-leg jacket with K-braces turns out to be the most inexpensive design, respectively followed by the thee-leg jacket, the optimised four-leg jacket and the reference design. It is expected that the four-leg jacket with K-braces brings total cost reduction of approximately nine percent compared to the reference design.

In the first fase of this study, attention is paid to the cross section of the dam for a plant in Cumberland Basin, capacity 1085 MW. Chosen is for generation over the ebb with use of bulb turbines. The main subject of this study was the design of the turbine caisson. The main subjects are Design principles, Dimensions, Loading and Foundation. The subvolume is about the enormous tidal difference (in dutch).

During the construction of the first Dutch offshore wind farm prototype measurements were performed. These measurements were aimed to monitor the behaviour of the granular filter layer of the scour protection around the mono-piles upon which the wind turbines are founded. These measurements were compared to scale model test results and theoretical analysis. The measurements show an overall lowering of the filter bed surface during the period that the filter beds were exposed to hydraulic loading of waves and tidal flow.

Nowadays, many modern countries are relying heavily on non-renewable resources. One common example of non-renewable resources is fossil fuel. Non-renewable resources are finite resources that will eventually dwindle, becoming too expensive or too environmentally damaging to retrieve. In contrast, renewable energy resources, like wind energy, are constantly replenished and are important because of the benefits it provides for us as well as for our environment. However, getting the energy from wind must be further investigated in order to make the usage of wind energy more economically. During the past thirty years, the trend in wind energy is to increase the size of wind turbines for producing more electricity power with lower cost. The increase in size is beneficial in terms of reduction of manufacturing costs per kW hours and reduction of the ground surface occupied by the wind farms. Further increases in size are not easily achievable because designers are expected to face more unknown technical problems such as aero-elastic stability problems. Therefore, it is important to investigate the aero-elastic stability problem of each new design concept in order to prevent the damage happen. In practice the larger blades have a lower edge-wise frequency that is closer to the flap-wise frequency than the case for the smaller blades. This could result in bigger edge-wise vibrations and unexpected aero-elastic problems. Larger blades will also result in large deformations even when the wind turbines are running at the design condition. Furthermore, designing pitch-regulated wind turbines will often result in lightweight and very flexible blades. The effects of large and flexible blade are mainly reducing diameter of the rotor during operation and coupling between edge-wise and torsional forces and motion. Most aero-elastic codes for wind turbines do not consider the effects of large deflections in their simulation of the loads and responses. The reduced effective rotor area leads to lower power production than predicted by linear calculations and the coupling between edge-wise and torsional forces and motion will increases pitch moment at the blade pitch system. The problems mentioned above are the current research problems needed to be investigated and solved due to upscaling of the wind turbines. In this research work, nonlinear flexible multi-body dynamics has been chosen to couple with nonlinear aerodynamics to investigate the current research problems mentioned before. An aero-elastic simulation tool called MBDyn-AeroDyn is developed by integration of two existing code named MBDyn [1] and AeroDyn [2]. MBDyn is a nonlinear flexible multi-body dynamic code, which can simulate the effect of large deflections and large rotations. The aerodynamic forces acting on wind turbine blades are calculated using AeroDyn which is based on the blade element-momentum theory. Improvement of aerodynamic calculation has been made by adding a modified Pitt-Peters dynamic inflow model in AeroDyn. Afterwards, both a linear time invariant system identification method and a linear time periodic system identification method have been used and implemented to investigate the aero-elastic stability of multi-MW wind turbine blades. Finally wind tunnel measurements have been performed in order to validate the aero-elastic simulation tool developed in this work. The validation of this aero-elastic simulation code has been carried out in this thesis. The time domain simulation results show that this aero-elastic simulation tool has good agreement with wind tunnel experiment results at the design operation condition. For the offdesign cases, the differences become larger gradually. Further more, an aero-elastic simulation has been applied on a 5MW wind turbine. Comparison has been made between FAST and BLADMODE in time domain. The results show that MBDyn-AeroDyn and FAST have a good agreement for an uniform wind field. Finally, flap-torsion flutter analysis on the 5MW wind turbine has been performed. The main feature of this aero-elastic simulation tool for multi-MW large horizontal axis wind turbines developed in this work is that kinematically large displacements and rotations are included, and that loads are applied on the deformed geometry. This allows the designers to simulate large wind turbines with more flexible blades to capture the effect of large displacements and rotations more accurately. Another feature is that both a linear time invariant system identification method and a linear time periodic system identification method have been implemented. It allows the designers to analyse stability from the time domain simulation data. From the result of flap-torsion flutter analysis on the 5MW wind turbine, it is shown that flap-torsion flutter is not likely to occur on this 5MW wind turbine with the current blade structural properties and intended operating conditions. Considering with the current Ph.D works which have been done so far, firstly, another wind tunnel experiment maybe necessary by using the same blades. Future measurement should be made at a higher rotation speed and combined with higher wind speed in order to find the stability boundary of this model wind turbine. Secondly, this aero-elastic simulation tool should be integrated into a knowledge based engineering(KBE) wind turbine blades design environment in order to take advantage of this aero-elastic simulation tool. Further research work should be carried out to simulate wind turbine blades with smart control devices, for example, flaps on the blades. Last but not the least, an user friendly graphical interface for this tool should be made in order to make it more convenient to use for a wider group of users.