This thesis is centered around the possibilities of integrating offshore wind power together with oil and gas platforms. The motivation behind this topic is to reduce the emissions of CO2 and other pollutive gases from conventional offshore power plants. The electrical systems on oil platforms are weaker than an onshore grid, so measures should be taken to let the wind power integration contribute to a more stable operation on the platform. To explore existing and future technologies that can achieve this is an important part of this work. Two different power systems are presented, denoted System 1 and System 2. Their schematics are shown in Fig. 1. Corresponding simulation models are built from fundamental blocks in the software PSCAD/EMTDC. Aggregated models are utilized in order to save computational time.System 1 consists of an offshore wind farm and an oil platform connected together in islanded operation. The oil platform contains an synchronous generator with an associated gas turbine and can adjust both active and reactive power quickly. The load consists of fixed-speed induction motors. The wind farm contains a back-to-back Voltage Source Converter (VSC) which is used for variable speed operation of the turbine. The converter is also used for voltage support to the system, and this functionality reduces voltage oscillations during disturbances. The most critical scenario investigated is to start a large induction motor. It is shown that the transmission cable may contribute to an increasing risk of voltage collapse during the start-up. Another critical event is when the wind power is suddenly disconnected, and the gas turbine has to adjust the power output quickly in order to avoid too large frequency deviations. The simulations show that a disconnection of the whole wind farm does not lead to critical operation or possible instability. This would not be the case in a system with slower control systems in the gas turbine and synchronous generator.System 2 consists of a VSC-HVDC connection between Kollsnes and the platform Troll A. This existing configuration is powering a gas compressor on the platform through a variable speed synchronous motor. This thesis proposes to integrate wind power on the DC-side of the Troll A VSC-terminal. The challenge is to inject wind power in a way such that the operation of the gas compressor is not disturbed. The proposed control system is working as intended, and the selected simulation cases show that the compressor system is not affected by the wind power. The DC voltage control system is able to maintain a constant voltage at the Troll terminal during normal operation. The wind farm reduces the losses in the HVDC-cable, and surplus wind power is sent to the land grid during low-load operation. The DC-voltage drops to 73 % of the rated value for 10 ms when the wind power is suddenly disconnected, but the duration is so short that it does not affect the gas compressor operation.The simulation results indicate that both configurations are feasible. However, the results are heavily dependent on the parameter data, and further research should put more efforts into gaining as correct values as possible. A sensitivity analysis is performed to System 1 as a guideline to which parameters that are most decisive, and therefore should be modeled most accurately. In addition, new simulation cases might reveal challenges that this thesis does not concern.

SIMPLE, SIMPLER, SIMPLEC and IDEAL are solution procedures originally developed for incompressible flows and staggered grids. For SIMPLE, SIMPLER and SIMPLEC, extensions for collocated grids and for treatment of flows at all speeds have already been proposed. For IDEAL, only an extension for collocated grids has been found, and an extension for treatment of flows at all speeds is proposed here. Extended versions of SIMPLE and SIMPLER are implemented in Brilliant, a multiphysics CFD-program developed by Petrell AS. These implemented algorithms are compared to the existing solution procedure in Brilliant, an extended version of the SIMPLEC algorithm. As expected, SIMPLE and SIMPLEC gave almost identical solutions for all the three presented test cases. The values given by the SIMPLER algorithm differed slightly from the values given by the two other algorithms. When simulating a shock tube, all three algorithms showed large deviations from the quasi-analytical solution in some regions of the shock tube. The SIMPLER algorithm spent the least CPU time for this simulation example, while SIMPLE and SIMPLEC spent less CPU time than SIMPLER when simulating methane flow in a pipe. Even though the CPU time was not registered for the last simulation example, a pressure relief pipe, it was noticed that the time consumption was much greater for the SIMPLER algorithm than for SIMPLE and SIMPLEC.

In this paper theory on cylinder and wind turbine wakes have been studied, and experimental work on the wake behind a wind turbine have been carried out in the Fluids engineering laboratory at NTNU.The objective of this paper is to show and explain how the wake from the tower of a wind turbine develops and interacts with the rotor wake. It is desirable to study the wake for different operating conditions of the wind turbine to see how the wake development is affected. A summary of classical wake theory, aerodynamics and wind turbine wakes will be given. Measurements in the wake of a cylinder fitted with pressure taps for drag calculation will be compared to theory and used as a reference. Also, the wake behind the wind turbine tower with the blades taken off will be studied and compared to the tower wake found behind the operating wind turbine.For comparison, reference measurements were done in the wake behind a cylinder and behind the free standing wind turbine tower without blades. The drag coefficient obtained from pressure measurements on the cylinder surface were 1.077 and match the expected value of 1.2 fairly well. However, neither the shape nor the maximum velocity deficit measured in the wake fit the theoretical profile. Drag coefficients calculated from the momentum deficit across the wake were only in the range of 0.65, which is almost half of the expected, and the huge deviation from theory could not be explained. With values between 1.07 and 1.50 the measured drag coefficients in the wake of the tower alone were also not consistent with theory. The shape of the tower wake profile coincides better with theory than the cylinder wake, but the maximum velocity deficit is generally lower than predicted by theory. Difference in drag can be explained with blockage effect and the smaller velocity deficit may be attributed to the free stream flow over the top of the tower interfering with the wake downstream of the tower.Wake surveys behind the wind turbine were done at three operating conditions: Optimum tip speed ratio; low tip speed ratio, with power output half of output at best point operation; and high tip speed ratio, with power output half of output at best point operation. The increased turbulence level behind the rotor the flow seen by the tower is believed to creates a turbulent boundary layer which stays attached to the surface to a point further back on the tower, creating a narrower and weaker wake compared the free standing tower wake. Optimum turbine operation gives a stronger rotation of the wake doe to the higher torque on the blades compared to the two other cases. At high TSR the wake is more uniform, and the tower wake disappears faster than in the wake of the turbine operating at lower TSR. The Strouhal number found in all the wakes match well with theory and does not seem to be affected by the rotor wake except that the tower vortices dies out quicker.

It is in this report given an analysis of the effect of utilizing load control in the Norwegian power system, which is one of the functions included in the Smart Grid consept. The analysis is performed by running simulations in the EMPS model. Effects are investigated in terms of increased socio-economic surplus in Norway and reduced prices during peak load periods in the area Østland.Based on earlier studies it is assumed that it is possible to disconnect 600 MW of electric water heaters during peak load hours without reducing any comfort for customers. In this analysis, 600 MW is shifted from two peak load hours to a low load period, implying that 6 GWh/week (600 MW*2h/day*5 days/week) is shifted from peak load periods to low load periods.Simulations using the existing model of the Nordic system show a net increase in socio-economic surplus of 20.208 MEUR per year by implementing a load shift. The average prices during the peak load periods are in this case reduced by 0.06 cent/kWh. In an extreme situation, a price reduction of 33.02 cent/kWh is observed. Including variation of wind power production in these simulations shows that the average price reductions are smaller and that the effect obtained in extreme situations is significantly smaller. This indicates that the effect of a load shift is somewhat smaller than what the results in the other simulations have shown.An improved system model, where the functions quadratic losses and gradual consumption adaption are included, give that a load shift increases socio-economic surplus by 41.198 MEUR per year. The average price reduction are found to be about the same as obtained by the original model. The price reductions in extreme situations are however considerably lower.Varying the exchange prices with the continental areas gives increased price differences between the price periods, which results in an even larger effect of load shift. The effect is especially large in a scenario where the prices in Germany and the Netherlands are very low during off-peak periods due to increased wind power production while the peak prices are high due to use of gas power plants instead of coal. An average peak price reduction of 0.07 cent/kWh and a price reduction of 10.48 cent/kWh during extreme situations are here found.Some uncertainty is connected to the results due to difficulties when comparing results from different simulations, which occurs when the calibrating the models. However, the results still indicate the range of the values that a load shift provides, which is clearly positive.

Gevinstene ved høy delta T i fjernvarmenett er velkjente; lavere varmetap, høyere virkningsgrad i varmeproduksjonsenheter, mindre pumpearbeid og større kapasitet i nettet. Hvor stor delta T i fjernvarmenettet blir, avhenger i stor grad av utformingen av abonnentsentralene og reguleringen av dem. Antall trinn varmeveksling i abonnentsentralen har innvirkning på total delta T, og ulike studier har her konkludert noe forskjellig. Enkelte har funnet at tre-trinns-sentraler gir best nedkjøling av primærvannet ved høye utetemperaturer, mens to-trinns-sentraler er best ved lave. Andre studier har funnet det teoretiske potensialet for nedkjøling til å være likt for to- og tre-trinns-sentraler.Underdimensjonering og tilsmussing av varmevekslere fører til lavere primær delta T. Seriekobling av radiatorsystem og ventilasjonssystem gir bedre nedkjøling av primærvannet. Akkumulatortanker i tappevannssystemet kan ha praktiske fordeler, men reduserer delta T.Riktig innstilling av ulike børverdier i tappevannssystemet er viktig for å oppnå høyest mulig delta T. Vannstrømmen i sirkulasjonsledningen bør være lavest mulig. For innstilling av alle parametre i tappevannsystemet må man samtidig passe på å overholde temperaturkrav i forhold til legionellasmitte. Optimalisering av vannstrøm og temperatur i radiatorsystem gir også en gevinst. Målinger er utført i en totrinns abonnentsentral tilknyttet fjernvarmenettet i Trondheim. I sentralen er det installert en tank mellom ettervarmeren og blandeventilen i tappevannssystemet for å dempe temperatursvingninger. To radiatorkretser og en ventilasjonskrets er koblet i parallell.Temperaturen ut fra blandeventilen bør være 65 °C for å holde temperaturen over 60 °C i hele sirkulasjonsledningen. Den sekundære temperaturen ut fra ettervarmeren bør være 72,5 °C for å holde 70 °C ut fra tanken. Dette gir en temperaturforskjell på 7,5 °C. Den målte temperaturen mellom disse målepunktene er hele 17 °C i gjennomsnitt. Dette fører til en større enn nødvendig innblanding av kaldt vann i blandeventilen, og lavere total delta T over sentralen. Fjernvarmeleverandørene krever en primær delta T på 50 °C over varmeveksleren for romoppvarming ved dimensjonerende utetemperatur(-19 °C i Trondheim). Målinger viser at dette kravet brytes allerede ved utetemperaturer rundt -5 °C og estimater peker mot en delta T på 30-33 °C ved dimensjonerende utetemperatur. Dette skyldes at varmeveksleren for romoppvarming er underdimensjonert.Systemvirkningsgraden er forholdet mellom levert energi til sluttmålet og den primære energien som trengs for å få det til. Primærenergifaktor for elektrisk energi er 3,31 mens den for varmeenergi levert til fjernvarmenettet i Trondheim er 1,46. Det er da antatt at søppel har en primærenergifaktor på 1.Systemvirkningsgraden for tappevannssystemet i den undersøkte abonnentsentralen er funnet til å være 0,2367. Det største varmetapet foregår i fordelingsnettet i bygget. 59 % av varmeenergien levert til abonnentsentralen går tapt her. Hvis man regner med at 25 % av den tapte energien kan nyttes til romoppvarming ender man opp med en systemvirkningsgrad på 0,3238.Beregninger viser at systemvirkningsgraden kun blir marginalt lavere ved å innføre akkumulering av tappevann i tappevannssystemet.

Erfaringer fra de seneste årene har vist at utfall av overføringskabler til utlandet og redusert produksjonskapasitet ved kjernekraftverk i Sverige, har stor betydning både for forsyningssikkerhet og prisdannelse i kraftmarkedet i Norge. Konsekvensene ble spesielt tydelige vinteren 2009/2010. Flere av kjernekraftverkene i Sverige gjennomgikk da omfattende revisjon- og vedlikeholdsarbeid sommeren 2009, og da arbeidet viste seg å være mer omfattende enn ventet, vedvarte perioden med redusert tilgjengelighet mye lengre enn ventet.Denne rapporten tar for seg tilfeller med ulik grad av reduksjon i tilgjengelig produksjonskapasitet ved kjernekraftverk i Sverige. Konsekvensene av redusert tilgjengelighet presenteres gjennom kategoriene kraftpris i markedet, utveksling, kraftproduksjon, magasinnivå og samfunnsøkonomisk overskudd. Samkjøringsmodellen er brukt for å gjennomføre simuleringene. Ved å redusere tilgjengeligheten av kjernekraft fra 98 % til 50 % under en periode på vinteren, øker den gjennomsnittlige spotprisen i Norge, Sverige og Nederland, både før, under og etter perioden. I en periode på 25 uker med redusert tilgjengelighet, når spotprisen i enkelte områder en gjennomsnittspris på over 1,5 Euro-cent/kWh mer enn under samme periode uten redusert tilgjengelighet.I tillegg til lengden på perioden med redusert tilgjengelighet, er det også av betydning om reduksjonen er ventet eller ikke. En uforutsett periode med redusert tilgjengelighet på 12 uker, som etterfølger en ventet periode på 13 uker, fører til ulike resultater, sammenlignet med om alle de 25 ukene med reduksjon hadde vært ventet. Ulikhetene kommer spesielt frem i resultater for magasinnivå og vannkraftproduksjon under perioden med redusert tilgjengelighet av kjernekraftverkene.Fordelingen av når på året vannkraft i Norge blir produsert påvirkes av en reduksjon i tilgjengelig kjernekraft i Sverige. For å kompensere for reduksjonen i kjernekraft, øker vannkraftproduksjonen under perioden med redusert tilgjengelighet. I ukene før og etter reduseres imidlertid produksjonen av vannkraft. Ved sammenfall av tørrår og redusert tilgjengelighet er det spesielt viktig å regulere mengden vannkraft som produseres i de ulike periodene av året.

Many applications benefit from using converters which can operate at high temperatures among them; down-hole drilling, hybrid vehicles and space craft. The theoretical performance of transistors made of Silicon Carbide (SiC) is investigated in this work. It is shown that their properties at high temperatures are superior compared to Silicon (Si) devices. Two half-bridge converters, using SiC normally-off Junction Field Effect Transistors (JFET) and SiC Bipolar Junction Transistors (BJT), are designed and tested to verify the performance of SiC devices. The challenges which arise when replacing slow switching Si devices with extremely fast SiC transistors are thoroughly discussed. To fully utilize the properties of SiC transistors, the converter designer must pay extra attention to the Printed Circuit Board (PCB) layout and component selection. For best device performance, all parasitics should be kept at a minimum. A high value of the Schottky diode capacitance results in voltage and current ringings, thus more ElectroMagnetic Interference (EMI) and switching energy loss. It is recommended selecting a Schottky diode with low junction capacitance as the freewheeling diode of the converter. Simulations and measurements show that the SiC transistors require special gate driving circuits to perform at their best. It is recommended to use a two-stage driver with high capability of charging and sinking gate/base currents. The JFET driver should also have a bipolar output to provide high noise immunity. The JFET is classified as a unipolar device; however the nature of the gate-source structure, which is a pn-junction, gives the device some bipolar characteristics. The gate structure is similar to BJT. To obtain a low on-state resistance some continuous gate current has to be provided. This is more pronounced at high temperatures.An analog control board, for operating the half-bridge as a step down converter, is designed and implemented. The controller uses current mode control to obtain a constant output voltage. Slope compensation is included to ensure stable operation at high duty-cycles. This makes the converter able to operate stably for a wide range of input voltages. The step down converter shows excellent performance during stationary operations and the ripple voltage is within its limit. When operating at high ambient temperatures it is important to have small transistor losses to minimize the self-heating of devices. The losses of a three phase inverter are calculated based on the measured conduction and switching losses of the SiC transistors. The calculations show that both inverters, based on BJTs and JFETs, are able to operate with very high efficiency even at high temperatures. High temperature characterization of the devices shows that their dynamic behavior, i.e. switching losses and switching times, are almost unaffected by temperature changes. The transistors, tested in this work, are suited to operate at ambient temperatures up to 150 ˚C. This is imposed by the conventional device packaging. Higher operating temperatures are achievable by using different packaging technologies, i.e. metal packaging advanced soldering methods.

The hydraulic models representing hydro turbines and conduit system found instandard model libraries of power system analysis tools are often simplied mod-els. Subsequently, important information about the dynamics of the hydraulicsystem may not be properly represented by such models, putatively resultingin insucient representation of the interaction between the electric system andhydraulic system.In this master thesis three dierent hydraulic models for hydro power plantsequipped with Francis turbines for use in power system simulation software hasbeen studied: 1) a simplied model often found in power system simulation tools;2) a model including a surge tank and elastic water column and 3) a model thatincludes a surge tank, elastic water column and turbine parameters accountingfor the characteristics of the hydraulic turbine.The hydraulic models were implemented in Simpow, a power system simulationtool. A frequency scan in the range from 10􀀀3-5 Hz was performed. The re-sults were compared with a frequency scan from LVtrans, a program specicallydesigned for accurate simulation of the dynamics of the hydraulic side in hydropower plants. The comparison showed that the simplied model failed to prop-erly represent the dynamics of the conduit system. The the model with surgetank and elastic water column was able to represent the dynamics of the con-duit system with satisfactory accuracy. Best representation was achieved for themodel including turbine parameters.The three hydraulic models were implemented in three dierent power systemcongurations: a single machine innite bus system; a system consisting of twointerconnected areas; and a system that has sustained power oscillations. Theresulting active power delivered from the generator were the hydraulic modelswas implemented, the speed of the turbine, the pressure at turbine and the owthrough the turbine were investigated.The simulation results revealed that the active power variation from the gener-iiators is in the same range for all three models, except for the simulation withsustained power oscillations. The speed variations of the turbine as a result ofincidents in the electrical network are in the same range for all three models.The model including turbine parameters is the only model able to representthe pressure variation as a result of a variation of speed of the turbine. Forpower oscillations with frequencies equal to the half period frequency of the waterhammer eect,1.38Hz, both the model with surge tank and elastic water columnand the turbine parameter model show very little response. For frequencies equalto the water hammer eect, 0.69Hz the variation in ow is also small for the twomodels. In general, the model with turbine parameters are better damped thanthe two other models.Further work should include development of an automated routine for determin-ing parameters to use in the model with turbine parameters as well as investiga-tions of how the model behaves in dierent network congurations.

This thesis work shows implementation and testing of existing control strategies in a novel offshore passive platform configuration. Control of the platform voltage and frequency is done through control of a shore-based B2B-VSC configuration, when the frequency on shore differs from the frequency on the platform. An 80 km HVAC-cable and transformers separates the shore-based B2B-VSCs from the platform installation. The system is modeled and simulated in MATLAB® Simulink™ SimPowerSystems™. The grid-side VSC is applied conventional cascaded controllers, with an inner current control loop and outer loops controlling the DC-link voltage and the flow of reactive power from the grid. Two different control strategies were implemented for the platform-side VSC, denoted as control strategy 1 and 2. Control strategy 1 controls the platform voltage magnitude, while keeping the frequency fixed, and is implemented with a simple PI-controller. Control strategy 2 controls both platform voltage magnitude and frequency, and was based on a cascaded control configuration, similar to that of a weak-grid system, implemented for the passive platform system. Both control strategies were implemented for the platform system and tested for three worst-case changes of platform load: loss of all loads, increase of passive load and large induction motor starting. The platform steady-state and transient voltages and frequency were evaluated based on the requirements for voltage and frequency defined in IEC 61892: standard for mobile and fixed offshore units, electrical installations. Control strategy 1 was tested for both normal and autonomous operation, when a constant DC-link voltage was applied. It was also tested for full B2B-VSC configuration, with the grid-side system VSC controlling the DC-link voltage and flow of reactive current from the grid. In all the three configurations, control strategy 1 failed to meet the system requirements set by the IEC 61892 standard for allowable voltage magnitude transients during loss-of-all-load and large-induction-motor-start. Control strategy 2 was tested with constant DC-link voltage when platform voltage measurements were available. The simulation results show no visible voltage or frequency transients for any of the three load change tests applied, and the system operation satisfies the IEC 61892 requirements for all the worst-case load change tests. The platform voltage and frequency are independent of the load dynamics, which is a desirable quality when controlling the voltage on a complex platform configuration. Control strategy 2 shows promising potential for controlling the platform voltage and frequency from shore. However, further research and testing must be done before this control strategy can be utilized in a real platform system. The proposed system controlled by control strategy 2 would allow removal of the low efficiency, high emission gas turbine driven synchronous generator that are the standard power source on oil platforms today. Eliminating the need for, or usage of the gas turbine, for power production on offshore oil and gas platforms, could reduce the total CO2 and NOx emissions of the petroleum sector considerably.

The metabolic insertion of titanium into the biosilica frustules of the pennate diatom Pinnularia sp. and the centric diatom Coscinodiscus wailesii was explored in the present study. A total of five titanium incorporation experiments were conducted on Pinnularia, in addition to five control experiments. Titanium dissolved in HCl was co-delivered with silicate dissolved in NaOH to silicate replete and silicate deplete Pinnularia cultures over a 10 hour delivery period. Cell number, silicate and titanium concentration, Quantum Yield, Instantaneous Chlorophyll Fluorescence and pH was measured regularly throughout the experiments. The diatoms were left to consume the delivered silicate/titanium for 62 hours after delivery was completed before they were harvested and rinsed to remove the organic material. The resulting biosilica frustules were analyzed with (Scanning) Transmission Electron Microscopy to investigate the frustule structure, Inductively Coupled Plasma Mass Spectrometry to measure the elemental composition of the frustules and Energy Dispersive Spectroscopy to determine the location of the elements in the frustules. Quantum Yield measurements were conducted to obtain information about the Pinnularia photo system condition, and high Quantum Yield values were interpreted as an indication of a good overall physiological state of the cell. It was found that a combination of high Quantum Yield values and silicate depleted cultures produced the highest titanium content in the Pinnularia frustules (1.3 wt% relative to silicon). Furthermore, EDS analysis revealed that titanium was not evenly distributed throughout the biosilica frustule. The relative titanium content in the Pinnularia frustules was higher in the biosilica at the base of the large pores covering the frustule surface than in the biosilica between the pores. This was in accordance with findings reported in literature. The observed inhomogeneous distribution trend was found to apply also to phosphorous and iron present in the titanium-doped frustules. This observation has not been previously reported.Cell aggregation of the Pinnularia cultures was observed upon silicate/titanium addition. Based on results from control experiments there was reason to believe that the aggregation was triggered by titanium, but the explanation for this cellular response as well as its effect on titanium uptake and incorporation was not determined. Three titanium incorporation experiments were conducted in the proof on concept study on Coscinodiscus . Titanium dissolved in HCl and silicate dissolved in NaOH was co-delivered with a pipette once a day for three, five and seven days. On the day following the final addition, the diatoms were harvested and rinsed before they were analyzed in the same manner as Pinnularia. The experiments on Coscinodiscus revealed that metabolic insertion of titanium into the frustule biosilica was possible, but due to the small maximal titanium content achieved (0.03 wt% relative to silicon) and the slow growth rate, Coscinodiscus was not considered an optimal choice of diatom species for titanium incorporation purposes.It was not possible to rule out whether the frustule structure in any of the two species had been altered or impaired as a consequence of titanium exposure. Structural flaws were observed in frustules with and without titanium incorporated. A faint irregular pattern on the nanoscale was observed in Pinnularia frustules containing titanium. No such pattern was observed in titanium free frustules.The amount of boron measured in the frustule biosilica was higher than the regular dopant concentration in boron-doped silicon. The boron content in Pinnularia was found to be on average 0.1 wt% or 0.28 atomic % relative to Si, while the average in Coscinodiscus was nearly twice as high. Without further manipulation, this high boron natural boron concentration makes diatom frustules unusable for doped semiconductor purposes.Future work should involve reproduction of the obtained results with more replicates, as well as larger culture volumes and, in the case of Coscinodiscus in particular, longer time scale. Titanium uptake and incorporation studies should if possible be performed on single Coscinodiscus cells. Multiple continuous silicate starvation and titanium/silicate delivery cycles should be tested to determine if it is possible to incorporate titanium into a larger fraction of the culture population, as well as to increase the maximum amount of titanium incorporated into one frustule.Aggregation of Pinnularia diatoms should be further investigated to determine reason for the observed reaction, the effects on titanium uptake/incorporation and possible strategies to prevent it from occurring. Pinnularia frustules unexposed to titanium should be inspected in order to determine if the iron and phosphorous distribution pattern observed in the titanium-doped frustules is present in titanium-free frustules as well.A completely artificial culture medium and plastic cultivation containers should be applied to explore how the boron content of diatom frustules can be tailored by controlled delivery of boron to the culture medium.