Submerged sediment-laden ice blocks that form on the intertidal mud flats of the Minas Basin pose a potential threat to tidal turbines planned for deployment in the Minas Passage. Laboratory prepared ice blocks of varying sediment content, salinity, and length scale were melted in seawater of different temperatures. The effect of sediment inclusions on melt rate is related to changes in heat supply and the heat required to melt a unit mass of ice, where the former is affected by the strength of the convective current and the latter by the ice block properties. A melt rate model for submerged sediment-laden ice is developed, with free convection the dominant deterioration mechanism. The model provides probable upper limits to the lifetimes of submerged ice blocks in the field, and has been used to predict lifetimes of large submerged ice blocks using temperatures representative of seawater in the Minas Basin.

Tidal stream energy presents challenges that will require the development of new engineering tools if designs are to harness this energy source effectively. At first glance one might imagine that tidal stream energy can be treated as wind with appropriate adjustment for fluid properties of water over air, and account taken of the harsher offshore environment; both waves and turbulence. However, it is now well accepted that the flow past turbines that are constrained by the local sea bed, sea surface, and possibly also neighbouring turbines and channel sides, will differ markedly from that of an ostensibly unblocked wind turbine. Garrett & Cummins (2007) were the first to demonstrate that operating a turbine in a non- negligibly blocked flow passage presents a different flow solution and importantly a significant opportunity to enhance the power that can be delivered by blocked turbines with the limit of power extraction exceeding the Lanchester-Betz limit for operation of unblocked wind turbines. Although it is impractical to array real turbines across the entire width of a channel it has been proposed to use short arrays of turbines making use of local constructive interference (blockage) effects; Nishino & Willden (2012) showed that although the phenomenal power limits of Garrett & Cummins are unobtainable in a real flow, a significant uplift in the limit of power extraction can be achieved for short fences of turbines arrayed normally to the flow in wide cross-section channels. However, it does not follow that rotors designed using unblocked wind turbine tools are capable of extracting any more power than they are designed for and hence the power uplift made available through blockage effects may be squandered. This thesis sets out to develop design tools to assist in the design of rotors in blocked environments that are designed to make use of the flow confinement effects and yield rotors capable of extracting some of the additional power on offer in blocked flow conditions. It is the pressure recovery condition used in wind turbine design that requires relaxation in blocked flow conditions and hence it is necessary to resort to a computational framework in which the free stream pressure drop can be properly accounted for. The tool of choice is a computational fluid dynamics embedded blade element method. As with all models with semi-empirical content it is necessary to select and test correction models that account for various simplifications inherent to the use of the blade element method over a fully blade resolved simulation. The thesis presents a rigorous comparison of the computational model with experimental data with the various correction methods employed. The tool is then used to design rotors, first for unblocked operation, with favourable comparison drawn to lifting line derived optimal Betz rotor solutions. The final objective of the study is to design rotors for operation in short fence configurations of four turbines arrayed normally to the flow. This is accomplished and it is shown that by using bespoke in situ rotor design it is possible to extract more power than possible with non-blockage designs. For the defined array layout and operating conditions, the bespoke rotor array design yields a power coefficient 26% greater than the implied Betz limit for an unblocked rotor and 4% greater than operating a rotor designed in isolation in the same array.

Before tidal stream energy is exploited, tidal power resource and environmental assessments must be undertaken. This thesis explores limits to power extraction for tidal sites defined by a strait between an island and landmass. Numerical simulations provided by Fluidity are used to analyse power extraction from locations in the strait and around the island for an idealised island-landmass domain and an actual coastal site. The numerical model is verified by comparing predictions with analytical solutions for inviscid flow past a circular cylinder located in the centre of a channel and in the vicinity of a wall. The model is then validated against laboratory measurements of flow patterns for impulsively-started flow past a submerged circular cylinder, and for flow past a surface-piercing circular cylinder in oscillatory laminar shallow flow. It is demonstrated that the numerical method captures satisfactorily the mechanisms of early wake formation, which indicates the model can be applied to assess tidal stream resource within the coastal geometries considered herein. Finally, the methodology to account for power extraction is satisfactorily verified for bounded and unbounded flows. Contrary to current practices, results from a parameter study for different idealised coastal sites reveal that the maximum power extracted in the strait is not well approximated by either the power extracted naturally at the seabed or the undisturbed kinetic power. Moreover, an analytical channel model underpredicts the maximum power extracted in the strait due to its inability to account for changes in the driving head resulting from power extraction and flow diversion offshore of the island. An exception is found for islands with large aspect ratios, with the larger dimension extending parallel to the landmass; i.e. the island-landmass geometry approaching that of a channel. In this case, the extracted power is satisfactorily approximated by the power naturally dissipated at the seabed and there is good agreement with the analytical model. The maximum power extracted in the strait is shown to decrease when water depths offshore are greater than in the strait, underlining the importance of fully understanding the wider bathymetry of a given site. A similar conclusion is reached when strait blockage is reduced. The power extraction in the strait is found to be sensitive to both viscosity and seabed friction, and these parameters need to be properly estimated during the setup and calibration of models in order to reduce uncertainty. Power extraction increases when turbines are sited simultaneously both in the strait and offshore. Tidal power assessment is performed for Rathlin Sound, off the coast of Northern Ireland. Again, no clear relationship is found between maximum power extracted in the strait and either the power dissipated naturally at the seabed or the undisturbed kinetic power. A similar ratio of power extracted to undisturbed kinetic power is obtained as for the equivalent idealised model. The analytical channel model underpredicts the maximum power extracted. The actual and idealised coastal site models indicate similar responses to changes in seabed friction, and similar reduction in power extraction with decreasing strait blockage.

Accurate characterisation of fluid flow at tidal energy sites is critical for cost effective Tidal Energy Converter (TEC) design in terms of efficiency and survivability. The standard instrumentation in tidal site characterisation has been Diverging acoustic-Beam Doppler Profilers (DBDPs) which remotely measure the flow over a range of scales, resolving up to three velocity vectors. However, they are understood to have several drawbacks particularly in terms of characterising turbulent aspects of the flow. This characterisation is generally based upon a small number of key transient metrics, the accuracy of which directly impacts TEC designs. This work presents an optimisation and performance assessment of newly available Single Beam Doppler Profilers (SBDPs) mounted on a commercial-scale tidal turbine at mid-channel depth in a real operating environment. It was hypothesised that SBDPs would have advantages over DBDPs for site characterisation, in terms of reduced random error, reduced uncertainty in turbulence intensities and the ability to quantify the structure of the turbulent flow. The relationship between random error, sensor orientation and flow speed is quantified for both single and diverging beam sensor types. Random error was found to increase with increasing flow velocity as a power law, the slope of which varies for different sensor orientations. Quantification of noise offers a practical method to correct turbulence metrics. To enable the use of multiple acoustic sensors mounted in close proximity, interference was quantified and mitigation techniques examined. Cross-talk between sensors of the same type were generally shown to bias measurements towards zero. In the presence of alternate types of acoustic sensors, interference caused an increase in standard deviation of velocity results. Implementing a timing offset control mechanism was able to mitigate this effect. This work has achieved a greater understanding of the drivers (spatial separation, inclination angle, pulse power) and effects on measurements of interference along with ambient-noise for users of acoustic instruments. Lessons learned of value to the industry, as site characterisation work intensifies ahead of next generation commercial scale devices, are presented. Mid-channel depth mounted SBDPs were found to have advantages over seabed mounted DBDPs in resolving the key turbulent flow metrics. SBDPs were able to resolve integral length-scales of turbulence that show an anisotropic ratio of scales as predicted from theory and in work at similar sites, while the DBDPs results were similar for all directions. Turbulence intensity measurements were found to be similar after noise correction, with the SBDPs more able to accurately capture the turbulence dissipation rate. This evidence suggests that SBDP arrays present a significant improvement over bottom mounted DBDPs in discerning information about the nature of the turbulent flow, and thus future site characterisation work should consider the use of SBDPs alongside bottom mounted DBDPs for this purpose.

Tidal energy is a clean, sustainable, reliable, predictable source of energy. Recent developments in underwater turbines have made harvesting tidal energy feasible. Determining the power potential available in a given water body can be accomplished by using numerical hydraulic models to predict the flow velocity at a location of interest. The East River in Manhattan has been used here in an effort to develop a modeling methodology for assessing the power potential of a site. Two two-dimensional CFD models, FESWMS and TUFLOW, as well as one one-dimensional model, HEC-RAS, are used to analyze flows in the East River. Comparisons are made between the models and TUFLOW proves to best represent flows in the East River. HEC-RAS provides accurate results; however, the one-dimensional results lack the necessary detail of a two-dimensional model. FESWMS cannot produce results that mimic actual flow conditions in the East River. Using the TUFLOW model, power and energy estimates are made. These estimates show that a two-dimensional model, such as TUFLOW, can be a great tool for engineers and planners developing tidal energy projects. Using the results of this work, a methodology is developed to assess power potential at other sites using publicly available data. / Master of Science

Στην παρούσα μεταπτυχιακή διατριβή προσπαθούμε να δώσουμε μια τάξη μεγέθους της παλιρροϊκής ενέργειας που υπάρχει στα παλιρροϊκά ρεύματα του Στενού Ρίου – Αντιρρίου και προτείνουμε την εγκατάσταση υδροστροβίλων “παλίρροιας” με την οποία θα μπορούσαμε να εξάγουμε την ενέργεια αυτή για να την εκμεταλλευτούμε ως ηλεκτρική ενέργεια στα σπίτια μας. Στο Κεφάλαιο 1, γίνεται μία εισαγωγή σχετικά με το φαινόμενο της παλίρροιας εξηγώντας τον τρόπο και τα αίτια που το δημιουργούν. Στο Κεφάλαιο 2, αναπτύσσονται κάποιες βασικές σχέσης υδραυλικής οι οποίες μας βοηθούν στον υπολογισμό της ταχύτητας και της ισχύς του παλιρροϊκού ρεύματος. Επίσης, γίνεται κατηγοριοποίηση των υδροστροβίλων ‘παλίρροιας’, καθώς παρουσιάζονται συγκεκριμένα μοντέλα εταιριών με τα τεχνικά χαρακτηριστικά τους. Στο Κεφάλαιο 3, περιγράφεται ο αλγόριθμος του οργανισμού EPRI, με τον οποίο γίνεται ο υπολογισμός της συνολικής μέσης ετήσιας παλιρροϊκής ενέργειας, η οποία μπορεί να είναι διαθέσιμη και να “εξαχθεί” από έναν υδατικό πόρο. Προτείνεται μια παραλλαγή της μεθόδου αυτής προσαρμοσμένη στα δικά μας δεδομένα και πραγματοποιείται σύγκριση μεταξύ των δύο. Στο Κεφάλαιο 4, παρουσιάζονται τα αριθμητικά και γραφικά αποτελέσματα για την ισχύ των παλιρροϊκών ρευμάτων του Στενού Ρίου – Αντιρρίου. Ορίζεται η περιοχή μελέτης και επικεντρωνόμαστε σε αυτή για την ανάλυση των χρονοσειρών της ταχύτητας των ρευμάτων. Στο Κεφάλαιο 5, παρουσιάζεται η εγκατάσταση τριών μοντέλων υδροστροβίλων στην περιοχή μελέτης. Τέλος, στο Κεφάλαιο 6, παρουσιάζεται μια περιβαλλοντική μελέτη για το έργο της εγκατάστασης υδροστροβίλων ‘παλίρροιας’ στο Στενό Ρίο – Αντιρρίου και παρατίθενται τα γενικά συμπεράσματα που προέκυψαν από τα προηγούμενα κεφαλαία. Παρατηρήσαμε, ότι η ισχύς του ρεύματος που μπορούμε να εξάγουμε από την περιοχή μελέτης που ορίσαμε είναι κατά μέση τιμή περίπου το 60% του μέγιστου επιτρεπτού που μπορούμε να εξάγουμε από τη συνολική διατομή του Στενού Ρίου - Αντιρρίου. Τα συστήματα των υδροστροβίλων, που προτείνονται μετά από την θεωρητική εφαρμογή τους στην περιοχή μελέτης, φαίνεται να μπορούν να καλύψουν με ηλεκτρική ενέργεια μεγάλο μέρος της περιοχής της Πάτρας. Οι δυσκολίες που θα έχουμε να αντιμετωπίσουμε στην περίπτωση της εγκατάστασης υδροστροβίλων “παλίρροιας” θα είναι ο τρόπος με τον οποίο θα τοποθετηθούν οι βάσεις των υδροστροβίλων στον πυθμένα, λόγω της σύστασης του πυθμένα, αλλά και οι αποστάσεις που θα πρέπει να τηρηθούν από τους πυλώνες της γέφυρας Ρίου – Αντιρρίου, το οποίο θα μπορούσε να λιγοστέψει τον αριθμό τον υδροστροβίλων που θα μπορούσαμε να εγκαταστήσουμε στο σύστημα. Τέλος, ένας σημαντικός παράγοντας που θα πρέπει να εξεταστεί είναι ότι ένα τέτοιο έργο μπορεί από περιβαλλοντικής άποψης να είναι θετικό, αλλά από οικονομικής πλευράς θα αύξανε αρκετά τους λογαριασμούς της Δ.Ε.Η, για λόγους συντηρήσεις, αρχικού κόστους εγκατάστασης, περιβαλλοντικών τελών κτλ. / -

Numerical and experimental investigations of tidal current energy extraction have been conducted in this study. A laboratory-scale water flume was simulated using commercial computational fluid dynamics (CFD) code FLUENT. In the numerical model, the tidal current turbine is represented with an actuator disk, which produces a pressure drop associated with energy loss. The free water surface is considered in the model using a volume of fluid method and is allowed to deform freely. Numerical results identified that a localised wake is formed behind the tidal current turbine and there is considerable localised flow acceleration around and most especially, under the energy extraction device. A free water surface drop is visualised in the model results due to the energy extraction and this free surface drop is believed to have an impact on the recovery of turbine wake. The influence of other parameters like water depth, ambient turbulence and flow speed on the tidal current energy extraction are also testified, based on the numerical model. Numerical results demonstrated that, because of the existence of a free water surface, tidal turbine interaction with the flow is a complicated three dimensional problem. Therefore, completely using the theoretical methods of wind turbines for tidal current turbine study would be inappropriate. Two physical tests were deigned for the experimental investigation of energy extraction from tidal currents and were carried out under different testing conditions: one was in moving water using a natural open channel and the other was in still water using a towing tank. Comparing experimental and numerical results of wake velocity profiles, good qualitative agreement has been obtained, which proves that the proposed numerical model can provide essential insight into the mechanism of wake development behind tidal current turbines. Experimental results also confirmed that, although moving water is the real operational condition of tidal turbines, a towing tank is still an ideal facility for the experimental study of tidal turbines, especially at the early stages of understanding of the detailed physical processes governing the performance of rotors and turbine wake behaviour. This study is a comprehensive investigation into tidal current energy extraction at laboratory scale. Environmental impact of tidal current energy extraction is further recognized and an appropriate experimental facility for the model testing of tidal energy extraction devices is recommended.

In recent years, the actuator disc approach which employs the Reynolds-Averaged Navier-Stokes (RANS) solvers has been extensively applied in wind and tidal energy field to estimate the wake of a horizontal axis turbine. This method is simpler to administer and requires moderate computational resources in modelling a tidal turbines rotor. Nonetheless, the use of actuator disc approximation in predicting the performance of tidal devices has been limited to studies involving an extremely small disc (e.g. rotor diameter of 0.1 meter). The drawback of a small scale actuator disc model is the overestimation of essential parameters such as the mesh density and the resolution of the vertical layers, making them impractical to be replicated in a regional scale model. Hence, this study aims to explore the methodology on implementation of the Three- Dimensional (3D) actuator disc-RANS model in an ocean scale simulation. Additionally, this study also aspires to examine the sensitivity of the applied momentum source term and its validity in representing full-size tidal devices. Nonetheless, before the effectiveness of an actuator disc in a regional model can be tested, tidal flow models for the area of interest needed to be set up first. This was essential for two reasons: (a) to ensure accurate hydrodynamic flow conditions at the deployment site were replicated, (b) to give confidence in the outputs produced by the regional scale actuator disc simulations, since in-situ turbine measurement data from a real deployment site were difficult to source. This research was undertaken in two stages; in the first stage, a numerical model which can simulate the tidal flow conditions of the deployment sites was constructed, and, in the second stage, the actuator disc method which is capable of modelling an array of real scale-sized tidal turbines rotors has been implemented. In the first stage, tidal flow simulations of the Pentland Firth and Orkney Waters (PFOW) were conducted using two distinct open-source software - Telemac3D, which is a finite element based numerical model, and Delft3D, which is a finite difference based model. Detailed methodologies in developing a 3D tidal flow model for the PFOW using both numerical models were presented, where their functionality, as well as limitations were explored. In the calibration and validation processes, both models demonstrated excellent comparison against the measured data. However, Telemac3D was selected for further modelling of the actuator disc considering the model's capability to perform parallel computing, together with its flexibility to combine both structured and unstructured mesh. In the second stage, to examine the actuator disc's accuracy in modelling a full size tidal device, the momentum source term was initially applied in an idealised channel study, where the presence of a 20-meter diameter turbine was simulated for both single and array configurations. The following parameters were investigated: (i) size of the unstructured mesh utilised in the computational domain, (ii) variation in disc's thickness, (iii) resolution of the imposed structured grid to represent turbine's enclosure, (iv) variation in the vertical layers, and (v) influence of hydrostatic and non-hydrostatic formulations on the models' outputs. It is to be noted that the turbine's support structures have not been included in the modelling. The predicted velocities and computed turbulence intensities from the models were compared against laboratory measurement data sourced from literature, where excellent agreement between the model outputs and the data from literature was observed. In essence, these studies highlighted the efficiency and robustness of the applied momentum source term in replicating the wake profiles and turbulence characteristics downstream of the disc, hence providing credence in implementing the actuator disc method for a regional scale application. Subsequently, the validated actuator disc method was applied to the Inner Sound region of the Pentland Firth to simulate arrays of up to 32 tidal turbine rotors. The wake development, flow interactions with the rotor arrays, and flow recovery at the Inner Sound region have been successfully mapped. Also, this study highlighted the importance of employing optimal numerical margins, specifically for the structured grid and horizontal planes, as both parameters were relevant in defining the disc's swept area. As published materials on the implementation of actuator disc approach within a regional scale model is still scarce, it was aspired that this work could provide some evidence, guidance and examples of suggested best practice in effort to fill the research gap in modelling tidal turbine arrays using the actuator disc approach.

As developers seek to convert the energy of the tides into electricity, sub-sea turbines must be designed to perform well in increasingly harsh conditions. Such energetic seas have historically been avoided, hence measurements taken below the surface in strong tidal currents and large waves are relatively few, and the theory behind these interactions is underdeveloped. This thesis compares measurements of subsurface velocity taken in the field, at a UK site proposed for development, to the velocity outputs of a model capable of combining waves and currents in a number of ways. In particular the interaction between waves and currents is investigated. The methodology incorporates a novel virtual velocity measurement instrument to measure the model flow, replicating the physical instruments used at sea, such that direct comparisons can be made between the two data-sets. Model and field velocities show good agreement across a range of current speeds and wave heights, with a range of metrics used to demonstrate the suitability of the model, based on linear wave-current theory, for this site. The wave-current interaction module is calibrated, with linear superposition of wave and current velocities proving a suitable representation of field velocities. Calculation of a dispersion relationship affected by mean current velocity marginally improves calibration with field data. Analysis of other sites using the tools developed will further validate this type of model, which in combination with blade element momentum theory, is able to predict pre-construction site specific loads on tidal turbines. Doppler Current Profilers (DCPs) are able to measure subsurface water particle kinematics and sea surface elevation simultaneously, however assumptions made by these instruments jeopardise detail when recording in highly energetic seas, particularly where waves and turbulent tidal currents combine. Models developed to optimise the design of tidal turbines require correct site specific inputs to accurately reflect the conditions that a turbine may encounter through its lifetime, moreover, the kinematics of these models must be accurately validated. To overcome the limitations in DCP measurements a 'Virtual' Doppler Current Profiler (VDCP) is developed (Crossley et al. 2017), enabling quantification of error in site characteristics, and 'like for like' comparisons of field and model kinematics that has never previously been documented. The numerical model developed incorporates tidal currents, waves and turbulence combined linearly to output subsurface velocity based on conditions from the field which have been averaged over ten minute intervals. The inputs are simple, time averaged characteristics (current magnitude, direction, and profile; wave height, period and direction, turbulence intensity and turbulence length-scale) and the model outputs velocities over a two dimensional grid that develops with time. The VDCP samples this flow as if it were the very instrument in the field that recorded the data used for validation. Taking into account the heading, pitch and roll of the instrument a data set directly comparable to that measured in the field is generated. The VDCP is initially used in quantifying error in wave and turbulence statistics, demonstrating a phase dependency of velocity measurements averaged between beams and providing a theoretical error for wave and turbulence characteristics sampled under a range of conditions, in order to improve tidal site characterisation. Spectral moments of the subsurface longitudinal wave orbital velocities recorded by the VDCP can be between 0.1 and 9 times those measured at a point for certain turbulent current conditions, turbulence intensity measurements may vary between 0.2 and 1.5 times the input value in low wave conditions and turbulence length scale calculations can vary by over ten times the input value, dependent on both current and wave conditions. The methodology can be used to determine a theoretical error in any site characterisation parameter for any set of wave, current and turbulence conditions. Results of the model validation using the VDCP show that the tidal flow model, and in particular the newly developed wave-current interaction module, is effective in simulating field subsurface velocities over a range of depths, for waves of up to 3m significant wave height and currents of up to 3.5ms-1. The model is effective in reproducing the wave climate using both measured and modelled surface elevation spectra, and tests, with marginal improvements, the effect of modifying the dispersion equation to account for currents. Field and model velocities compare well over the frequency range dominated by waves, showing only small underestimations in model standard deviations with respect to those from field data, at depths close to the sea surface. At the low frequency end of the modelled spectra, where large turbulent eddies dominate, there is some deviation in model accuracy, particularly during the ebb tide where recorded turbulence parameters are extremely variable, creating uncertainty due to a relatively small sample size. Between field and model velocity maxima, some scatter is observed, potentially providing uncertainty in the estimation of ultimate loads. Model and field damage equivalent velocities, used in the determination of fatigue loads, agree well. Results suggest that a linear wave-current representation of subsurface velocities at this particular tidal site is applicable. Care should be taken when interpreting this result due to the relatively small sample size, and the possibility of site specific nuances, and as such further studies are proposed. The Virtual DCP model is a novel development which has proven its usefulness in the work contained in this thesis and in the analysis of commercial field data. It is extremely versatile, adapting to a range of configurations and set up criteria such that it can be used in the quantification of DCP measurement error for a range of flow characteristics. This information is useful in the design of tidal turbines (and other sub-sea structures) as well as for oceanographic and biological processes. The tidal flow model developed extends beyond the capability of similar numerical models with the capability to model the interaction between waves and currents according to a number of different options. Combined with the VDCP, which samples from the model flow field, a system is created that can be effectively calibrated to find the best model solution to replicate flows at a tidal site measured by a 'real' DCP over a broad range of sea conditions and water depths. The purpose is to ensure that models used to predict the sub surface velocities in the field are suitable and a key question was to understand whether the linear super-position of linear wave models and a turbulent current flow provides a realistic model of the particle kinematics with a view to undertaking loads analysis of a tidal stream turbine. Comparisons of this kind have not previously been documented, and this thesis lays out the path to improved site characterisation.

Nowadays, renewable energy is in full growth. In particular, offshore wind farms will be at the centre of UK energetic strategy in the coming years. However, other types of marine renewable are still at an early development stage. That is the case for tidal energy. Many projects have been undertaken but there is no candidate for competitive commercial applications yet. Deltastream is one of these numerous pioneering projects. It consists of a set of three marine current turbines mounted on a triangular base put down onto the seabed. The device is not moored and no harm is done to the environment. However, that makes the structure more sensitive to water flows. And it is important to ensure that it will remain at its location and not being carried along with the tidal streams. Using CFD, the present study aims to evaluate the drag on the nacelles of the structure and come up with solutions to reduce it as much as possible.