AIM (Advanced Instability Methods) for industry

Lead Research Organisation: University of Cambridge
Department Name: Engineering

Abstract

In the last ten years there has been a surge of interest in non-modal analysis applied to standard problems in fundamental fluid mechanics. Even in simple flows, the behaviour predicted by these non-modal analyses can be completely different from - and far more accurate than - that predicted by conventional analyses, particularly for the types of flows found in industrial situations.The successful application of non-modal analysis to standard problems sets the scene for step changes in engineering practice. Nevertheless, some very significant challenges must be overcome. Firstly, the standard approach cannot handle the non-linear problems often found in engineering. Secondly, the standard approach is computationally expensive and cannot handle problems with many degrees of freedom. Thirdly, the standard approach deals with simple measures, such as kinetic energy density, while other measures are usually more pertinent for industrial situations. Encouragingly, applied mathematicians and engineers have made significant progress in all of these areas. This progress has revealed that a generalized formulation of the problem in terms of constrained optimization and variational methods, adapting and applying methods from the control and computational communities, will bridge the gap between standard flows and engineering problems.Our vision is that future generations of engineering Computational Fluid Dynamics (CFD) tools will contain modules that can perform non-modal analysis. If and when such analyses can be made practicable they are certain to change the way that engineers design fluid mechanical systems, such as combustion chambers, turbine blades, reaction chambers and ink jet printers. Furthermore, they can readily deal with transient effects and non-periodic time-varying base flows, which are often particularly relevant in engineering situations.This research will benefit UK industries that rely on the modelling and control of fluid mechanics and thermoacoustics. For example, the pharmaceutical industry will benefit from a better understanding of transition to turbulence and relaminarization in physiological flows, which is important for the application of drugs via the nose and upper airways; The gas turbine industry will benefit from being able to perform instant sensitivity analyses of their fuel injectors and to combine this with greater understanding of the thermo-acoustics that leads to combustion instability; and the wind turbine industry will benefit from an improved prediction of the sensitivity of an aerofoil to turbulence transition and results of exposure to a gust or to the wake of the preceding aerofoil.The investigators in this proposal are all founder members of the EPSRC-funded Advanced Instability Methods (AIM) Network, which was set up in January 2009 to explore the relevance of non-normal analysis to industrial problems. Through masterclasses and workshops in academia and industry and an increasing number of web-based resources, the network provides a route for dissemination and exploitation of this research.In summary, the objectives of this proposal are to bridge the gap between fundamental work and engineering practice, to embed these techniques in the engineering design cycle and to reinforce a growing centre of excellence within the UK in this area. The generalized framework proposed here, combined with two challenging engineering examples and the resources of the AIM Network, will make this possible and demonstrate it to a wider engineering community.

Planned Impact

This research will benefit the commercial private sector in the UK, particularly industries that rely on modelling and control of fluid mechanics and thermoacoustics. For example: the pharmaceutical industry will benefit from a better understanding of transition to turbulence in physiological flows; the gas turbine industry will benefit from sensitivity analyses of their fuel injectors and from non-modal analysis of combustion instability; the industrial combustion, energy, chemical processing and pharmaceutical industries will benefit from improved capability to predict mixing rates and eventually pollutant formation; commercial CFD users will benefit from stability tools in post-processors and optimization routines; the inkjet printer industry will benefit from the potential to create non-spherical droplets; and the wind turbine industry will benefit from an improved understanding of the sensitivity of an aerofoil to turbulence transition. This research will increase the standard of technology and the global competitiveness of UK-based companies. It will accelerate investigation of new problems and provide new insight into current problems. This will lead to new industrial design tools that will help high tech UK industries to remain competitive in markets in which technological advantage plays a crucial commercial role. As university researchers, we can create a core technology with a wide range of applications, which would not normally be funded by a single company. This will provide benefits within three years through new insights into current problems. We have found that that our research has greatest impact when we interact with industrial code development teams as they adopt our research to the format that is most useful to them. This process will take three to ten years. In the long term, we anticipate that industry will invest over a ten to twenty year timescale, employing some of the original developers of the technique. In this project, researchers across four departments will have to tackle un-solved problems, interpret research from a variety of backgrounds and translate it into a format that can be easily understood by the industrial community. These activities will develop problem solving, communication and teamwork skills. The investigators are founder members of the EPSRC Advanced Instability Methods (AIM) Network. The core technology will be made available to all members - currently over 70 academic and 7 industrial. We have already developed tutorials, handouts and model numerical code that can be downloaded and adapted by members. We will augment this with industrial case studies, In the longer term, the core technology will be useful to other companies with similar questions. As part of the network, we are organizing workshops targeted at particular problems across a range of industries. We will approach other potential partners directly and use the Research Days at Imperial College and Masters projects students at Cambridge to start collaborations. The investigators have their own established links with industry. We have regular (usually quarterly) meetings with our industrial contacts and will use these as opportunities to showcase research, to explore new areas and to indentify potential impact. The individual investigators will take the lead for interaction with their own industrial contacts, bringing experience from the other investigators as required. The post doctoral researchers will be involved and, if appropriate, the PhD students will be placed within industry for a few months. There will be no IP related to the concept of the generalized framework itself because this is already freely available. IP may arise from the specific application of the concept. Any collaborative work with the above companies will be performed under existing agreements, such as those between Cambridge and Rolls-Royce and BP.

Publications

10 25 50
publication icon
Arratia C (2013) Transient perturbation growth in time-dependent mixing layers in Journal of Fluid Mechanics

publication icon
Foures D (2013) Localization of flow structures using -norm optimization in Journal of Fluid Mechanics

publication icon
Foures DP (2012) Variational framework for flow optimization using seminorm constraints. in Physical review. E, Statistical, nonlinear, and soft matter physics

publication icon
Grimble T (2017) Local linear stability analysis of cyclone separators in Journal of Fluid Mechanics

publication icon
Illingworth S (2013) Finding thermoacoustic limit cycles for a ducted Burke-Schumann flame in Proceedings of the Combustion Institute

publication icon
Illingworth, S (2014) Acoustic state space models using a wave-based approach in 21st International Conference on Sound and Vibration

publication icon
Kashinath K (2013) Nonlinear Phenomena in Thermoacoustic Systems With Premixed Flames in Journal of Engineering for Gas Turbines and Power

 
Description We have transferred developments in non-modal stability analysis from the applied maths community to the engineering community. These techniques span a large class of engineering problems.

In combustion instability, one of the application areas of the research, we have checked the results experimentally and see how to embed the process within the design cycle. This will be done in future research projects.

One objective of the grant was to enable the UK to become the centre for the engineering application of non-modal stability analysis. There were many publications and conference presentations that arose from this project, as intended. Even better for this objective, one of the main architects of the application of this analysis to fluid mechanics, who was a partner on the proposal, moved from overseas to the UK during the grant.

Regarding the technical objectives, we successfully developed and tested the generalized framework for non-modal stability analysis and applied it to the two situations envisaged.

We have developed stability analysis methods for complex geometry flows which are now available in an open access software package, Nektar++. Not only have we provided methods to analyse two dimensional (biglobal) geometries using classical eigenvalue analysis as well as singular value decomposition we have also extended these techniques to so called Triglobal problems where one can analyse the instability pas a three-dimensional flow. Using triglobal analysis we have investigated how spanwise blowing and sucking can suppress vortex shedding of flow past a circular cylinder. Whilst the analysis is at low Reynolds numbers the instability persist to higher Reynolds numbers and so has potential broader application.

We have developed a mathematically rigorous method to determine the best way to mix fluids with different properties with a fixed amount of energy input. This method can be applied to a wide range of engineering problems, in a range of flow geometries.

Now that we have developed the DAL method for mixing problems, it can be applied to a wide range of situations. Possible future directions include optimising mixing in fluids with non-newtonian rheologies, flows in complicated geometries, or reacting flows, where the mixing also has a dynamical effect.
Exploitation Route The generalized framework developed during this research was applied to the problem of combustion instability. A simple model in Matlab is available online (http://www2.eng.cam.ac.uk/~mpj1001/papers/Rijketool_V06.zip). It can be downloaded and adapted by other researchers. This is most relevant to Rolls-Royce. The principles are clear, but further work is required in order to turn this into an industrial tool. This work is continuing in other research projects.

The tools developed under this grant have been used internally in Master student project and also as part of a related PhD project. We have also had a number of requests from external European groups who have been using the software.
Sectors Aerospace, Defence and Marine,Manufacturing, including Industrial Biotechology

URL http://www2.eng.cam.ac.uk/~mpj1001/MJ_biography.html
 
Description One aspect of the research was to convert a model used in industry into a format that could be treated with the generalised framework. This allowed the generalised framework for the study of flow stability to be applied to combustion instability problems. This will provide a systematic way to improve the design of gas turbine combustion chambers. This work is continuing as part of other research proposals with collaborators in the USA and France. The generalise framework is also being trialled in other industries, such as particle separation and carbon nanotube production. This has led to one funded iCASE student and one £2.7m research proposal (ANAM) recently awarded by EPSRC Developed and implemented two and three-dimensional asymptotic and transient stability analysis for complex geometry flows and made these methods available in an open-ware package. We have further used these tools to analyse triglobal stability of flow past a circular cylinder subject to span-wise blowing and sucking. We have demonstrated for the first time that the "Direct-Adjoint-Looping" (DAL) method can be applied to fully nonlinear problems where the optimising functional is not a norm. This means that this powerful method for identifying "optimal" initial conditions or flow forcing can now be applied to a wide range of problems of engineering relevance. We have also demonstrated that optimal mixing of a passive scalar in a fluid is not always achieved by maximising the kinetic energy of the perturbation.
First Year Of Impact 2013
Sector Aerospace, Defence and Marine,Energy,Environment,Manufacturing, including Industrial Biotechology
 
Description Council of Science and Technology Review on Modelling
Geographic Reach National 
Policy Influence Type Membership of a guideline committee
 
Description EPSRC Case Studentship
Amount £54,000 (GBP)
Organisation Dyson 
Sector Private
Country United Kingdom
Start 10/2015 
End 09/2019
 
Description Marie Curie Actions
Amount £243,000 (GBP)
Organisation European Commission 
Sector Public
Country European Union (EU)
Start 01/2012 
End 12/2015
 
Title Nektar++ version 4.0.1 
Description Nektar++ is a tensor product based finite element package designed to allow one to construct efficient classical low polynomial order h-type solvers (where h is the size of the finite element) as well as higher p-order piecewise polynomial order solvers. 
Type Of Technology Software 
Year Produced 2014 
Open Source License? Yes  
Impact The software is being used by a number of national and international groups and our web site is currently being visited up to 100 times a day according to google analytics 
URL http://www.nektar.info/downloads/file/nektar-source-tar-gz-2/
 
Title Rijketool 
Description A model of a thermoacoustic network on which nonlinear adjoint looping (the generalised framework) is applied 
Type Of Technology Software 
Year Produced 2012 
Impact This is an education tool. It runs alongside an academic paper on the subject 
URL http://www2.eng.cam.ac.uk/~mpj1001/papers/Rijketool_V06.zip