Liquid-liquid coalescers are devices used for increasing the droplet size of the dispersed phase in continuous phase flow, such as oil droplets in water flow. The efficiency of separation technologies is strongly dependent on the droplet size, which is desirable to shift into larger droplet diameters. Theory behind coalescence and its modeling is studied in this Maser’s thesis. Aker Process Systems AS, Division of Advanced Separation Technology, provided the assignment proposal.The scope of this work is a literature study on the coalescence phenomenon and the closely related break-up phenomenon and CFD modeling in general. Further a mathematical model for simulating coalescence of oil droplets in continuous water flow is developed by the use of the commercial CFD-code FLUENT. The basis for the model is a swirl-based coalescer called Compact Tubular Coalescer (CTC), developed by Aker Process Systems AS.The validity of the model is evaluated before different aspects of the performance of the coalescer are studied. Several validation criteria were tested and were acceptable, but some weaknesses regarding lack of test cases were detected. The performance testing showed good performance of the CTC, it was able to increase the Sauter Mean Diameter (SMD) of the droplet with up to 250% for the smallest droplets (20 μm) and highest volume fractions (7%). Remarkable differences of the performance were observed as the physical properties were changed. Higher viscosity and droplet surface tension lead to increased coalescence rate and decreased break-up rate.Future work is recommended to concentrate on improving the present model and to investigate more aspects of the model. An effort should also be made to use a Eulerian approach to model the dispersed phase with the use of population balances, in order to be able to simulate flows with larger dispersed phase volume fractions.

This thesis is on the design of an airfoil for a downwind wind turbine rotor with thin flexible wings, for offshore floating conditions. It has been suggested that such a system would to be lighter, simpler and allow for the use of more efficient airfoils. There has been a significant amount of work done at NTNU to develop a “high-lift” airfoil. These are airfoils with very high lift-to-drag ratios. They operate very efficiently at their design angle, but tend to not work well over a range of angles and conditions, and have a sudden and dramatic stall characteristic. In this thesis, it is attempted to pick up the work done with the high-lift profiles at NTNU in the 1980’s, and develop a new profile which has performance in the high-lift range, but with a much smoother stall and more stable characteristics, and to do so for the typical conditions expected for the suggested turbine. A fictitious 5 MW version of the suggested turbine was created and analyzed with the blade element momentum method (BEM). This gave informative results about the conditions the new airfoil must operate in. The high-lift technology and the earlier reports from NTNU were studied. Based on this knowledge and the numerical values from the BEM calculations, a serious of new airfoils were developed. By using the simulation programs Xfoil and Fluent (CFD), it was possible to modify and test a large number of airfoils and find the desired qualities.It was possible to design airfoils that had performance in the high-lift range, while maintaining stable operation and having a soft stall, and also increase the lift coefficient to be able to design for lower angles of attack. The profiles created here appear to be suitable for wind turbines, and provide an impressive increase in performance compared to traditional airfoils.Extra effort was put into making airfoils that were unaffected by roughness, air properties and Reynolds number, as stable performance in varying conditions are necessary for wind turbine blades. This was done by using adverse pressure gradients to control the point of transition.A slow stall was achieved by letting the pressure recovery distribution gradually approach the local ideal Stratford distribution when moving back over the airfoil. This caused the flow separate at the back first, and then the separation would grow gradually forward with increasing angle of attack.The inclusion of a separation ramp also worked very well together with the high-lift design, and allowed for an increased lift coefficient and more stable operation during the region of early stall.The most successful profile created appears to be the AR profile. It combines a diverged Stratford distribution with a separation ramp and a pressure spike at the nose to control transition. It has a wider range, stalls later and softer, and has a much more stable performance with varying conditions compared to the original HOG profile from NTNU. At the design point, the maximum performance is reduced only 5.9 % compared to the HOG. For higher and lower angles of attack, and increased values of roughness and turbulence, the AR has an all round higher performance than the HOG. It appears to be usable for wind turbines, and would increase the maximum airfoil performance by up to 40 % compared to commonly used NACA profiles. More good profiles were made, with varying thickness, stall and performance. Depending on the exact local requirements of an application, this report offers several interesting profiles to choose from. For instance, the D2 profile has round shape and over 16 % thickness, it has an even softer stall than the conventional wind turbine profiles, and would increase the maximum airfoil performance by up to ~34%. This profile would also be usable for upwind turbines.It was found that there is a big potential for manipulating the high-lift technology to give various shapes and performances. The usability of these profiles therefore appears to be wider than previously assumed.

Denne oppgaven har sett på modellering av spredning og fordamping av dammer på faste underlag bestående av kryogene væsker. En modell basert på grunt vann-ligningene har blitt utledet. Grunt vann-ligningene består av: én kontinuitetsligning og to impulsligninger i henholdsvis x- og y-retning. Modellen tar høyde for varmeovergang, avdamping og todimensjonal spredning av dammen. I tillegg inkluderes viskøse krefter som påvirker spredningen. Ligningene som oppstår fra modelleringen løses numerisk. For å gjøre dette har det blitt skrevet et numerisk program i Fortran 90. Programmet har blitt kalt CRYOSPREAD og de numeriske teknikkene dette programmet bygger på gås gjennom i teksten. Det har også blitt gjennomført simuleringer med dette programmet. Simuleringene sammenlignes med henholdsvis forsøksdata og modeller i Kameleon FirEx (KFX). Ved å sammenligne med forsøksdata fra spredning av vann, vises det at dam-spredningsmodellen i CRYOSPREAD er kapabel til å følge forsøksdataene rimelig bra. Dette er imidlertid svært avhengig av en turbulenskonstant som representerer mye av usikkerheten i modellen. Sammenligning med spredningsmodellen av dammer i KFX viser tydelig forbedringen ved å løse impulsligninger. Et simuleringsforsøk med begge modellene har blitt gjennomført, der strømningen i starten skal være dominerende i én retning. Resultatet viser at dammen i KFX-modellen med én gang sprer seg mer radielt og ikke har en dominerende retning for spredningen til forskjell fra den nye modellen i CRYOSPREAD. Forsøksdata for avdampingsrater av metan og nitrogen uten spredning viser at avdampingsmodellen i CRYOSPREAD presterer rimelig bra for senere tider (t > ~20s), men har noe avvik i starten. Sammenligning med avdamping i KFX uten spredning viser den kvalitative forskjellen i avdampingsratene som følger av forskjellige modeller for varmeoverføring. Total avdampet masse fra dammen i KFX er proporsjonal med tiden, mens den er proporsjonal med t1/2 i CRYOSPREAD. Forskjellen består i at KFX ikke tar hensyn til kjøling av underlaget, mens modellen i CRYOSPREAD tar hensyn til dette gjennom en vertikal konduksjonsmodell i bakken. Da det ikke lyktes å få tak i forsøksdata for spredning og avdamping av kryogener er ikke modellen tilstrekkelig validert, men vil kunne gi en god pekepinn for hva som skjer ved spredningsulykker av kryogener hvis antagelsene den bygger på er gode nok.

A literature study is first presented then the practical experiments are described. The theoretical chapter provides an overview of filtration mechanisms and available gas cleaning technologies, with an emphasis on hot gas and cake filtration. The experimental chapters describe the setup and testing of the filter tray test rig and the L10-56 test rig.The filter tray test rig was built, instrumentation was installed and programming performed. Puff-back calibration and heating tests were performed. The puff-back calibration showed that a significantly lower delta p minimum was obtained with this design compared to previous designs. Filtration tests were conducted at an elevated temperature (120 C) with different dust concentrations (1 and 3 g/m3) and filtration velocities (10 - 20 cm/s). The results showed that the rig was working in accordance with the design criteria, with a filtration efficiency above 99,5%.A previously used filter test rig, the L10-56 test rig, was installed at the slipstream of the exhaust of a oil sand extraction pilot plant at 350⁰C. A horizontal bed unit (HBU) was also developed to investigate the impacts of the exhaust gas on the granular filter media. Once completed, the rig was re-tested and the puff-back calibration and heating tests that were performed achieved results corresponding with the previous testis performed on this test rig. The test rig was ready to start filtration tests but they were not initiated due to time limitations and unpredicted halt of the oil sand extraction pilot plant.

Effekt av overflateruhet på diffusorstrømningMasteroppgave av Roy-Andre Midtgård, med veileder professor Per-Åge Krogstad.Formålet med dette prosjektet er å studere effekten av overflateruhet på et turbulent grensesjikt utsatt for en motvirkende trykkgradient. For å gjøre dette studeres det nedre grensesjiktet i en diffusorstrømning. To strømningstilfeller undersøkes og sammenlignes. I det ene tilfellet er bunnen av diffusoren glatt og jevn. I det andre tilfellet er bunnen dekket med periodiske ruhetselementer. Arbeidsverktøyet i dette studiet er numeriske beregningsverktøy, CFD. Programmene ANSYS Gambit/Fluent benyttes [1].Ved bruk av turbulensmodellene k-ϵ, k-ϵ RNG, standard k-ω og k-ω SST er det funnet konsistente resultater. Disse sammenlignes med forsøksresultater fra dr. Ing rapporten til Per Egil Skåre [2]. Simuleringsresultatene korrelerer til dels godt med forsøksresultatene, men de oppfyller i motsetning til Skåre ikke kravene til likevekt av turbulent produksjon og dissipasjon i grensesjiktet. Turbulensmodellene k-ϵ RNG og k-ω SST benyttes på det andre strømningstilfellet, og resultatene vurderes mot resultatene fra det første strømningstilfellet. Simuleringsresultatene fra k-ω SST indikerer en global separasjon av strømningen i grensesjiktet. Simuleringsresultatene fra k-ϵ RNG indikerer imidlertid ikke global separasjon. Skalert med indre variable finnes stigningstallet i det overlappende området til 0.54, altså 29 % høyere en von Kàrmàn konstanten på 0.418. Forflyttningen ned langs u+-aksen finnes til verdier mellom 12 og 15. Felles for simuleringsresultatene i strømningstilfelle 2 er det at ingen av dem oppfyller kravene til likevekt. Noe av årsaken til de nevnte avvikene mellom turbulensmodellene kan skyldes tilnærmelser i turbulensmodellene eller numerisk diffusjon. Det er også en mulighet for at trykkgradienten i diffusoren ikke er sterk nok til å danne likevekt gjennom måleområdet.

This thesis deals with the application of morphing skin and structure technology in a hydroturbine context. The work has been divided into two parts; construction of a demonstration rig and CFD calculations assessing the concept of camber morphing of guide vanes.Five rubber 'morphing bumps' reduce the intensity of vortex shedding behind a cylinder in an open flow of water. The bumps are made by a rubber bellows placed in a perforated stainless steel pipe. Their size is controlled by air pressure, and the option of reducing them to a neutral state is present. A control software has been designed using Labview, allowing manual and automatic (model predicted flow control) operation, as well as logging of measurements. The tests show that the bumps are capable of reducing vortex intensity by approximately 25%.CFD simulations has been performed on a Francis turbine section to determine if altering the camber of guide vanes at different loads can have a beneficial effect to the performance and wear characteristics. The simulations resulted in a dataset consisting of steady state and transient simulation results from 28 combinations of camber and wicket gate opening. The results did not establish that camber morphing can increase efficiency, but neither did they reject it completely. It has been shown that for small cambers up to 2%, the gains from increased lift far outweigh increased drag at part loads.

The main dynamic loads on a tidal turbine are due to the tidal current variation, turbulence, wave-current interaction and wake from the tower and upstream turbines. The design of the turbine is highly dependent on the dynamic loads acting on the turbine's blades and its structure. Cyclic loading and unloading of any material over a period of time will lead to fatigue and structural damage. In this work dimensioning loads for NTNU's 1 MW reference tidal turbine are calculated due to the vertical no-slip current velocity profile, wave-current interaction, the horizontal and vertical velocity component due to the waves, turbulence and the tidal wave. Based on mathematical models the dimensioning loads for the turbine are calculated with a Matlab code written for this work. The influence of each of the components causing a dynamic load on the tidal turbine blades is evaluated and analyzed with the rainflow counting algorithm, which can be used for fatigue analysis.Results from Matlab calculations are compared to experimental results in order to validate the Matlab code. In the comparison the trends are the same for calculation and experimental results, oscillating curves with the same period. It is noticed that in large waves the Matlab code predicts very well the dynamic loads in terms of mean and peak-to-peak values.For the thrust force and the shaft torque the horizontal component of the velocity due to the waves has the largest influence on the peak-to-peak values and the turbulence induces a high number of cycles. For the pitch moment as well the horizontal component due to the waves has the largest influence on the peak-to-peak values, but also the vertical component has a certain influence. For the pitch moment the consistently main oscillations inducing cyclic loading are due to the vertical no-slip current velocity profile. This is also the case for the yaw moment, but here the vertical wave component has the largest influence on the peak-to-peak values. For calculation with the tidal wave it is noticed that higher velocities gives higher maximum values. The Matlab code is a reliable and fast tool. Despite the simplifications done in the calculations, it calculates the dimensioning loads for a tidal turbine with a good accuracy.

This master's thesis is a continuation of previous theses written at ComputIT AS. It treats heat transfer to LNG pools boiling on water through two heat transfer models, LNGSIM1 and LNGSIM2. LNGSIM1 utilizes heat transfer correlations for pure liquids in combination with physical data of the mixture, while LNGSIM2 uses LNGSIM1 and a simple model for the concentration boundary layer. Both models are implemented in the CFD software Kameleon FireEx (KFX) and thereafter tested and validated against experimental data from the Burro test series. Comparisons with experimental data show that LNGSIM1 often produces correct trends in the downstream gas concentrations. The results are, however, often shifted in time, indicating that the heat transfer in the beginning of the spill is too low. LNGSIM2 is constructed to increase the heat transfer compared to LNGSIM1, hence vaporizing the LNG faster to better fit the experimental data in time. The choice of the constant CSIM2 in LNGSIM2 greatly affects the heat transfer, and it is found to fit experimental data best for 0.70 < CSIM2 < 0.80.An attempt to approximate LNG as pure methane produced erroneous results due to the heat flux remaining constant throughout the spill. Another attempted approximation was the use of a constant heat transfer coefficient. This produced very low heat fluxes towards the end of the spill, making it impossible for the gas concentrations to reach a zero value within the experimental time interval. The use of these simplifications are therefore not advised.A study of rapid phase transitions (RPT) is conducted using a simple criterion for when an RPT can occur. Comparison with a theoretical study gives promising results for describing when, where and why an RPT occur. This can be used to estimate when to release the pressure wave of an RPT.Investigations of the pool boiling correlations for pure liquids conclude that the way of calculating the transition boiling regime results in too high heat fluxes in that regime. To address this, a parameter study using LNGSIM1 and a factor Z is performed in order to reduce the transition boiling heat fluxes. The optimum values of Z are thereafter combined with the optimum values of CSIM2. Combining Z and CSIM2 reveals that most of the investigated values of Z overrides the wanted effect of CSIM2, hence warranting new approaches to reduce the overestimated transition boiling heat fluxes. Nevertheless, LNGSIM1 and LNGSIM2 with 0.70 < CSIM2 < 1.00 appear to be good alternatives to the current heat transfer model in KFX, since the heat transfer coefficient is continuously calculated based on compositions and boiling regimes, whereas the KFX model requires a constant heat transfer coefficient as input.

I september 2009 tok KLP Eiendom i bruk kontorbygget i Professor Brochs gate 2 (PB2) i Teknobyen i Trondheim. Denne bygningen har klimaskjerm og tekniske løsninger som tilsier langt lavere energibruk enn dagens forskriftskrav. Romoppvarmingen i PB2 er kun basert på fjernvarme, med gulvvarme i deler av 1. etasje og radiatorer i resten av bygget. Romkjøling gjøres kun via ventilasjonsanlegget med en kombinert kjølemaskin/varmepumpe tilkoblet kjøle-/varmebatteri i ventilasjonsaggregatene. Et annet mekanisk kjøleanlegg med tørrkjølere på tak dekker IT- og serverrom. Anlegg for snøsmelt dekker ca. 400m2 ved PB2 der fjernvarme også er energikilden.Ventilasjonen i bygget dekkes av 8 ventilasjonsaggregater der alle anleggene er behovstyrt etter tilstedeværelse eller CO2. Ventilasjonsvifter og VAV spjeld styres etter såkalte optimizersystemer med optimal styring av tilluftstrykk. Ventilasjonsluftoppvarming gjøres i sekvens der høyeffektiv roterende varmegjenvinner dekker første oppvarmingen, et kombinert varme- og kjølebatteri tilkoblet luft-vann varmepumpe tar videre oppvarming. Er ikke dette tilstrekkelig finnes et andre varmebatteri tilkoblet fjernvarme som tar resterende luftoppvarming. PB2 har et omfattende EOS og SD-anlegg med totalt 72 målere. Likevel finnes ikke måler på fjernvarmeforbruk til snøsmeltanlegget. Tilgjengelig energibruksdata i EOS database har til dels dårlig kvalitet med hele 23% av 2010 uten registrerte data. Dette er fordelt på både lange og kortvarige perioder. I tillegg viser flere av fjernvarmemålerne direkte feil forbruk i perioder med meget lavt forbruk.Spesifikk levert energi til PB2 var i 2010 på 115,5 kWh/m2år som etter energimerkeordningen gir energikarakter B. Dette forbruket er fordelt på 68,9 kWh/m2år elektrisitet og 46,6 kWh/m2år fjernvarme. Forbruket er høyere enn målsetningen på 94 kWh/m2år. I tillegg skal det her nevnes at bygningen ikke har vært fult utleid i 2010 slik at forbruket kan antas å øke ytterligere. På grunn av den dårlige kvaliteten på tilgjengelige energidata fra undermålerene ble det utført en korrigeringsprosess på disse dataene som bedret kvalitet på data med 20 %, likevel antas disse dataene å ha en feil på mellom 8 og 24% også etter korrigeringen. Ut fra disse dataene fantes spesifikk levert energi til romoppvarming til 27,7 kWh/m2år, ventilasjons oppvarming til 12 kWh/m2år, tappevann til 3,3 kWh/m2år, vifter og pumper til 7,4 kWh/m2år, belysning til 16,4 kWh/m2år, teknisk utstyr til 23,6 kWh/m2år samt kjøling til 9,4 kWh/m2år. Det er hovedsakelig romoppvarmingen som er betydelig høyere enn målsetning.Angående drivende faktorer for energibruk i PB2 er følgende konkludert: Tilstedeværelse driver elektrisitetsforbruk til belysning og teknisk utstyr, samt ventilasjonsvifter og tappevann. Utetemperaturen driver fjernvarmeforbruket til radiatorer, ventilasjonsaggregater samt gulvvarme. I tillegg er varmepumpeanlegg 35.01 påvirket av utetemperaturen med lavest forbruk rundt 12C, og økende ved høyere og lavere temperaturer. Brukeratferd og antall person som jobber i bygget har også stor innvirkning på energibruk til vifter, serverrom, utstyr og tappevann.Betraktninger av innetemperaturer på tre utvalgte kontorer i PB2 viser tilfredsstillende innetemperatur på alle tre kontorer i perioden. Variasjonen mellom hver av kontorene er til dels stor men dette antas å være på grunn av muligheten for individuell regulering av temperaturen på hver cellekontor. Likevel skal det nevnes at kontor 1 har en ugunstig temperaturfordeling i perioden med tanke på energibruk, der temperaturer er lavere i perioder med kjølebehov enn i perioder med varmebehov. Det er identifisert en del forhold ved PB2 som ikke er optimale løsninger. Det kan her nevnes lav utnytting av kondensatorvarme fra kjøleanlegg, unødvendig høyt gulvvarmeforbruk, unødvendig bruk av fjernvarme og varme fra varmepumpe i varmebatterier i ventilasjonsaggregater, samt relativt høyt energiforbruk utenfor driftstid til belysning og teknisk utstyr.Skulle PB2 blitt bygd med tanke på målsetning om nullutslipp vil tiltak som lavtemperatur varmeanlegg med akkumulatortanker tilkoblet solfanger og vann-vann varmepumpe kunne vært en aktuell løsning. I tillegg anbefales bruk av passiv kjøling primært og frikjøling med varmepumpebrønnene som varmesluk sekundært. Ventilasjonsluftoppvarming burde dekkes av høyeffektive roterende varmegjenvinnere primært, samt varmebatteri tilkoblet akkumulatortanker i varmeanlegget sekundært. Tilstedeværelse og dagslyskorrigert styring av høyeffektive lyskilder ville minimert forbruket til belysning.Skal nullutslippsbygg kunne bli en realitet i kalde klima som i Norge vil behovet for videreutvikling av eksisterende og nye løsninger være stort. Det kan her spesielt nevnes utvikling og mer kjennskap til høyeffektive varmegjenvinnere, forenklede lavtemperatur vannbårne systemer inkludert akkumulatortank, luftbåren oppvarming og frikjøling via ventilasjon, samt bruk av kulvert i hybride systemer for passiv forvarming/kjøling av ventilasjonsluft. Robuste og stabile styringssystemer er også helt essensielt her for å oppnå optimalt samspill mellom de ulike systemene i bygget. For å redusere behovet til elektrisitet til belysning og teknisk utstyr til et minimum er både meget energieffektivt utstyr, gode styringsprinsipper og en holdningsendring blant brukerne nødvendig.

This thesis presents two designs of tension-leg-platforms (TLP) support structures for the 10 MW reference wind turbine being developed by the Norwegian Research Centre for Offshore Wind Technology (NOWITECH). The designs result from iterative design processes which account for important design considerations such as performance requirements, natural frequencies and main cost drivers, and differ in their capability of providing stability to the wind turbine. TLP Towed is stable during towing and operation, whereas TLP Transported only provides stability when installed and is dependent on alternative transportation methods. The design processes are validated by investigating the influence from the various requirements and the sensitivity to wind turbine properties. The two resulting designs are compared and discussed in terms of cost competitive advantage. Fully coupled time-domain aero-hydro-servo-elastic models are established in FAST by using hydrodynamic computations from WAMIT, and the models are verified by comparisons to previous time-domain results and frequency-domain calculations. The natural frequencies of the FOWTs are obtained by model linearizations, and a discussion regarding overlap with wind turbine operational frequencies and wave excitation frequencies leads to modifications to the preliminary designs. A number of simulations with different wind and wave conditions are run and the TLP designs are compared based on displacements, upwind and downwind tether tensions, the nacelle's velocity and acceleration and extreme events. Resonant behavior, damping and instabilities are also discussed and suggestions for improvements to the designs are presented. The results presented in this thesis serve as guidance in the process of developing optimized TLP designs for an offshore wind turbine.