Experimental investigation of effects of section narrowing on solitary waves and boundary layer

N-Waves, which are a type of solitary waves, do not occur in the nature frequently so that the researchers can not examine them easily. Solitary waves, particularly N-Waves, are of particular importance in geosciences due especially to tsunamis generated by sea bottom landslides and earthquakes. Investigation of solitary waves is crucial in understanding natural phenomenon such as non-linear waves in nearshore regions or catastrophic disasters like tsunamis which has devastating effects. In the early years of solitary wave investigations, scientists decided that tsunami waves can be represented by a solitary wave (i.e. with only one crest). This decision led to deeper research on solitary waves, substantial amount of theoretical approaches followed by numerical and physical modelling studies. However due to unpredictability of tsunamis and lack of widespread visual recording devices, clear and well founded studies were not achieved in simulation of a real tsunami wave. Due to the recent dramatic advancement in digital technology, remote sensing techniques became widespread and also video and photo capturing technologies improved so advanced that each individual can have a camera of their own. Therefore, observation, our most basic scientific tool, gave us the ability to improve our knowledge on tsunamis. At the end of 90s, a trough accompanied by a crest was discovered in tsunami waves. The geometrical similarity of this wave and Latin letter "N" ended up with the name N-Wave. Additionally, two new definitions were made for N-Waves; if the crest is followed by a trough, 'Leading Elevation N-wave (LEN)' and if the trough is followed by a crest, 'Leading Depression N-wave (LDN)'. In the following years researchers developed different theories and numerical models, but due to lack of real world tsunami records and lack of tsunami type wave generators in laboratories, they had problems in verification of their theories and models in an accurate extent. In 26th of December 2004, a devastating tsunami formed during Sumatra Earthquake in Indian Ocean. During this disaster more than 230000 people from 14 countries lost their lives and incidentally a Belgium boat called 'Mercator' was able to take a record of water level variation during this tsunami. After this recording N-Wave theories were reconsidered, renewed and better established. According to the up-to-date studies, N-Waves have been considered as the independent superposition of two solitary waves in different phases; one has positive and the other has negative amplitude. This new studies are still in the beginning level with rare publications and N-Waves are being generated in laboratories just recently. In this context, an experimental setup was designed the laboratory flume of Istanbul Technical University Hydraulics laboratory such that a typical N-Wave was generated and investigated experimentally in the frame of this PhD thesis. Hydraulic Laboratory of Istanbul Technical University has a unique N-Wave generator with a special vertically-oriented piston mechanism in a wave flume with dimensions 23.5m x 1.0m x 0.5m. Previously a number of researchers investigated run-up characteristics of this kind of waves in this flume. Also, effects of coastal forests on tsunami run-up and coastal erosion, behaviour of rubble-mound and caisson type breakwaters under tsunami waves were investigated in this flume. In this experimental study, geometrical and kinematical properties of N-Waves were investigated to fill the gap in the literature. Accordingly, studies were conducted under two main chapters. Firstly, boundary layer measurements on the flume bottom were performed for better understanding of N-wave generated near-bed flow properties such as time-averaged velocity, turbulence and bed shear stresses. Afterwards as the second chapter, a simplified flume narrowing section was manufactured for better understanding the effects of coastal morphology (such as estuaries, straits, bays etc.) on the progression of N-waves. Therefore, changes in physical properties of investigated N-wave due to narrowing were presented. Furthermore, water elevation time series and velocity measurements were obtained. Resistant type probes were used for water elevation measurements. For this system two resistant type wave probes, one 8 channel HR Wallingford wave monitor and a four channel National Instruments NI-9215 A/D converter were used. Also one channel of the A/D converter was assigned for synchronization between water level measurements and velocity measurements. In other words, synchronization output of the velocity profiler was connected to one channel of the A/D converter in order to ensure simultaneous data recording. Velocity measurements were performed by using the 3D acoustic velocity profiler (Vectorino) manufactured by Nortek AS. This new generation profiler can measure simultaneously three component of the instantaneous velocity in 35 points over a vertical range of 3.5cm, with intervals of 1mm. Moreover, this profiler can perform with a sampling rate up to the 100Hz. During the measurements, this device sends short sound impulses to the flow and listens with receivers placed on its four arm. Consecutively, it compares transmitted and received frequencies of sound and calculates velocity of the flow by using the Doppler effect theory. This real time and simultaneous velocity data was recorded by using a program provided by the manufacturer of the instrument. This type of velocitymeters are widely used in coastal and hydraulic applications and proved their reliability through the years. For generating controlled conditions, a metal sheet with dimensions of 2m (length) x 1m (width) x 0.004m (thickness) was deployed at bottom of the flume in the measurement section. Also this metal sheet was painted to assure a smooth surface. All velocity measurements were performed over the same point on this metal sheet during this study. Thus, surface roughness (corresponding to hydraulically smooth conditions) was kept constant during the experiments. Furthermore, two measurement points were used at section narrowed experiments. To ensure identical bottom roughness properties in consecutive narrowing configurations, side panels used for section narrowing were slid instead of sliding measurement probes. Since N-Waves are unsteady and non-periodic by their nature, time-averaging techniques cannot be used to calculate the mean flow parameters. Therefore, ensemble averaging method was used during the investigation of its physical properties. In this context the identical N-Wave was generated for 40 times in successive experiments and velocity profiles close to the bottom and water elevation time series were measured simultaneously. Additionally, velocity profiles at different vertical ranges from the flume bottom were measured in order to determine free stream velocity for the same N-Wave. These measurements were analysed numerically by the Matlab software tool and results were presented in this thesis. In the light of the findings, the results of water surface profile were matched with N-wave equation with a high positive correlation. Thus, it can be said that produced wave in the context of this study is a typical N-wave. Afterwards free stream velocity, time-average velocity profiles, turbulent fluctuations and wave Reynolds number values were calculated. In the light of the findings, it was concluded that flow was in transitional state from laminar to turbulent regime at the phase of maximum free stream velocity. Also, flow was dominated by the horizontal velocity component (u) and the vertical velocity component (w) was very small. The results further show that the turbulent intensity and turbulent kinetic energy were also dominated by the horizontal component. Besides, turbulence was produced within the boundary layer close to bed and dissipated towards to upper levels. Moreover, it can be said that there is a net longitudinal mass transport in the flow field beneath N-waves. Also, a phase shift between water elevation time series and mean horizontal velocity was not observed, as expected in the case of solitary wave. Additionally, the relationship between free stream velocity and water elevation time series was investigated. It is found that velocities at rising curve of the N-wave were lower than the falling curve. This result is in contradiction with flood wave (gradually varied flow case), another type of unsteady flow. Furthermore, bottom shear stresses were obtained during the experiments. Total bottom shear stress consists of two parts which are viscous shear stress and turbulent shear stress. It was revealed that turbulent stress component was more dominant than the viscous stress counterpart, when the free stream velocity was at its maximum. Furthermore, fluctuations at the shear stress are important because of formation of vortices and turbulent coherent structures. Hence, fluctuation of the shear stress was also presented in this study. Tsunamis, first generated at offshore, are extensively affected by the coastal morphologies as they approach the shoreline. To investigate N-Waves reaching estuaries and river mouths, the second chapter of the experimental set-up was established. In this set-up, the width of the flume was reduced gradually with 5, 10 and 15 degree angles by means of a 2cm thick water-resistant plywood panels. Velocity profiles and water level variations were also measured for these 3 configrations. These measurements were made at two sections which were at the entrance of narrowed section and middle of the that section. In these measurement a significant phase shift was observed at the tail section of the N-wave. Also wave height and period differences were obtained and presented herein. Additionally, velocity differences were studied between narrowed section and un-narrowed section experiments. In this case, free stream velocities at narrowed section experiments were approximately 50% less than the un-narrowed velocities. This is an expected result for a general wave. Experimental measurements, for 70 experiments in total, were analysed in Matlab environment with codes written especially for this purpose. Flow properties like velocity profiles, shear stresses, turbulence properties were produces as results of these analyses. Graphs and videos of the results were given in Appendices. Since the process is highly unsteady, animation videos presenting the time-varying velocity and turbulence measurements are very indicative in terms of the characteristics of N-Wave boundary layers. Additionally, the effect of narrowing and water level variations were also included in the aforementioned visual material.