The application of modal decomposition techniques for the analysis of environmental flows

Civil Engineering and fluid mechanics are two fields that are not usually put into the same category. In reality however, many of the problems faced in Civil Engineering are caused by fluids. For example: the mixing of pollutants; the scouring of riverbanks; and the undesired oscillations of bridges and buildings caused by wind. Whilst it is possible to simulate the interactions of fluids and structures, it is computationally expensive and as a consequence experiments are needed to validate simplified models. Undertaking these experiments is not trivial, and unavoidably, the data collected can contain outlier / anomalous data points. There have been many attempts to create algorithms to automatically remove and or replace these outliers, the most effective or which are based on modal decomposition techniques. In fluid mechanics there are two prominent modal decomposition techniques. These are Proper Orthogonal Decomposition (POD), which can be use to determine modes that are spatially independent, and Dynamic Mode Decomposition (DMD) which can be used to determine modes that are temporally independent.
The majority of the previous POD and DMD applications have been have been applied to mechanical and aerospace engineering problems. However, the focus of this thesis is the application of modal decomposition techniques solely in Civil Engineering. First, the modal decomposition technique POD, is used to create a novel computationally efficient method for filtering spurious points from experi- mental data. Second a method of combining POD and DMD to attain regions of spatial and temporal independence is proposed. Third, these methods are applied to a river groyne problem to explain the spatio-temporal mechanisms leading to the sudden expansion of a mixing layer. Finally, these methods are applied to a
group of multi-scale square cylinders, resembling the layout of a city, to describe the spatio-temporal behaviour of the wake.
This thesis creates a suite of tools which can be applied by Civil Engineers to understand complex mechanisms, for instance, in environmental fluid mechanics.