Extracting coherent structures and low-order modelling using Optimal Mode Decomposition

Since nearly 25% of UK emissions are generated by transport, improving the energy efficiency of aviation and road vehicles is vital if the UK is to meet legally binding targets to significantly reduce greenhouse gas emissions by 2050. One promising approach to achieving this aim is to improve the aerodynamic performance of ground and air vehicles using active flow control. In active flow control, measurements are taken from the flow in real time and, in response to those measurements, action is taken to positively modify the flow. For example, movable surfaces on an aircraft's wing or on the rear of a road vehicle may be dynamically adjusted to reduce aerodynamic drag and hence decrease fuel consumption.

To implement an active control strategy typically requires an approximating flow model, which allows a beneficial control action to be computed from a given flow measurement. However, a significant challenge is that flows of practical interest have highly complicated and intricate dynamics, while computational and theoretical restrictions mean that the approximating models must be simple, or low-dimensional. One method of bridging this gap is to identify coherent structures in a flow, that is, spatial features which are of dynamical importance to its evolution and performance. If a small set of representative coherent structures can be identified, they can be used as the building blocks for a flow model and therefore facilitate a successful flow control strategy.

The aim of this project is to develop the Optimal Mode Decomposition (OMD) algorithm, a method which systematically extracts dynamically important coherent structures from ensembles of experimental or numerical fluid flow data. The systematic nature of the data extraction is important since it provides a tractable method of analysing the increasing large-scale sets of fluid flow data that are now available. The proposed research will thoroughly benchmark the performance of the OMD algorithm across a range of fundamental and challenging fluid flows, assess the suitability of the extracted structures to be used for flow modelling and estimation and, finally, undertake a theoretical study to explain any observed behaviour. This research will therefore develop a methodology to underpin the application of flow control for improved energy efficiency and consequently addresses a current and fundamentally important environmental and economic issue.

The transportation sector generates roughly 20% of EU countries' CO2 emissions. Although emissions from other industrial sectors, such as energy generation, are expected to decrease over the coming decades, a predicted increase in demand means this is not the case for transportation. Consequently, it is of great importance to improve the fuel efficiency of, for example, road vehicles and aviation in order to address key climate change targets. Since aerodynamic performance is fundamental to the fuel and energy efficiency of road and air vehicles, understanding and controlling the behaviour of fluid flows presents an important, but technically challenging, pathway towards achieving this aim.

This research seeks to address this challenge by developing a methodology - the Optimal Mode Decomposition algorithm - that will help researchers transition from the ability to accurately measure or observe a fluid flow, to being able to model, predict and improve its behaviour. Therefore, the impact of this research will be to produce a tool that enables the application of flow control technologies that can be used to improve fuel efficiency in transportation. This promises economic benefits through reduced fuel consumption and associated cost savings. More importantly, improvements in fuel efficiency will have environmental and societal benefits ranging from direct improvements to air quality to the vital medium-term impact of helping to mitigate the effects of climate change. Outcomes of the proposed research will be presented at an industry-sponsored workshop towards to highlight the developed methodologies outside of the academic community, and also at the Imperial College Festival, a major public outreach event, in order to motivate flow modelling and control as a fascinating and practically important challenge to the next generation of researchers.