Computational Toolbox for Fluid-Membrane Interaction with Applications to Micro Air Vehicles and Insect Flight
Lead Research Organisation:
University of Cambridge
Department Name: Engineering
Abstract
This research addresses the need for computational tools and techniques for the aeroelastic analysis of fluid-membrane systems. The particular focus is on development of a comprehensive computational toolbox for analysing the dynamics of fluids and embedded structures with the goal of obtaining high-fidelity predictions for aerodynamic design parameters, like drag and lift, and structural design parameters, like maximum stress and deflection. Although the equations for viscous, incompressible fluid flow and membrane dynamics are straightforward and well known, it is still challenging, if not impossible, to perform fluid and membrane coupled computations in the presence of large deformations. Besides fundamental differences in the mathematical structure of fluid and solid equations, progress is hampered by the vast disparity of the physical length and time scales involved. To address the first issue, the computational toolbox will include a mathematically rigorous and algorithmically robust formulation for representing the coupled dynamics of a fluid with an immersed membrane. For resolving the length and time scales a two-pronged approach will be followed. First, the developed techniques will be scalable to large system-level, three-dimensional simulations, with up to billions of unknowns, which will be achieved through systematic utilization of high-performance computing platforms. Second, the influence of the unresolvable sub-grid scales on the large-scale motions of the fluid flow will be explicitly modelled with a multi-scale method. The design of flapping wing micro air vehicles (MAVs) has been chosen as the driving application for this research. Conventional aerodynamic methods are either inapplicable or too crude for the design space exploration of bio-inspired MAVs with highly compliant flapping wings. Therefore, further MAV development requires new computational tools for the aeroelastic analysis of fluid-membrane systems. In return, the MAVs will provide an unrivalled testbed for a comprehensive experimental validation programme for the computational predictions. Flapping wing MAVs are nevertheless poor imitations of the natural millimetre and centimetre scale flyers - the insects. A combination of computations and free flight experiments will therefore be conducted towards identifying the key factors in natural flyers excellence. The highly scalable and validated computational toolbox developed during the project will be made available as open source software to the scientific community. Therefore, it is expected that the outcome of this project will be relevant far beyond the design of MAVs and the study of insect flight.
Publications
Bandara K
(2015)
Boundary element based multiresolution shape optimisation in electrostatics
in Journal of Computational Physics
Bandara K
(2016)
Shape optimisation with multiresolution subdivision surfaces and immersed finite elements
in Computer Methods in Applied Mechanics and Engineering
Bornemann P
(2013)
A subdivision-based implementation of the hierarchical b-spline finite element method
in Computer Methods in Applied Mechanics and Engineering
Cirak F
(2015)
Isogeometric Analysis and Applications 2014
Cirak F
(2011)
Subdivision shells with exact boundary control and non-manifold geometry
in International Journal for Numerical Methods in Engineering
De Clercq K
(2009)
Aerodynamic Experiments on DelFly II: Unsteady Lift Enhancement
in International Journal of Micro Air Vehicles
Elimelech Y
(2013)
Analysis of the transitional flow field over a fixed hummingbird wing.
in The Journal of experimental biology
Floreano D
(2009)
Flying Insects and Robots
Kolár V
(2013)
Average Corotation of Line Segments Near a Point and Vortex Identification
in AIAA Journal
| Description | Within this project we have developed a fundamentally new immersed finite element technique in which the fluid and structure meshes need not be matching. The fluid grid is fixed throughout the problem domain and the structure mesh is allowed to move freely. The developed approach is particularly appealing for applications with very large structural deformations. In contrast to conventional approaches, there is no need for error-prone mesh generation or re-meshing. In the developed immersed technique the fluid equations are discretized using a Cartesian grid to achieve algorithmic simplicity and scalability. The Cartesian grid also facilitates the use of smooth b-spline basis functions. It is becoming increasingly evident that b-splines are more accurate than the Lagrange basis functions used in almost all common academic and commercial finite element software. As an added benefit, b-splines enable an elegant new technique, referred to as subdivision stabilisation, for avoiding oscillations in the computed pressure field. In the implemented fluid-structure coupling technique the transmission conditions between the fluid and structure are enforced weakly. The resulting discrete systems of equations are solved with a partitioned approach. In each time step the fluid is solved with the known structure velocities as a Dirichlet boundary condition and subsequently the structure is solved with the known fluid tractions and velocities as a Robin boundary condition. The final algorithm is free of the frequently reported added-mass instability in fluid-structure coupling. In addition, we developed structural beam and shell finite elements as continuation of our earlier work on subdivision shells. This need arose from modeling of insect wings composed of thick veins and thin membranes. The new shear-flexible subdivision shell finite elements are well suited for thin as well as thick structures. They are inherently free from transverse shear locking effects and, thus, solve a problem to which literally thousands of papers have been devoted. The developed computational toolbox, named openFTL (open Finite Element Template Libray), has been validated with in-house and published experimental data on insect flight and animal locomotion in water. Amongst others, we simulated the leading vortex formation in drosophila wings rotating in a propeller apparatus in Reynolds number range 200 to 5000. The finely resolved simulations with in the order of one hundred million unknowns have been performed on HECToR, with run times less than eight hours. To obtain validation data, we developed a 'flapper' that accurately mimics the wing motion of insects. It can fly either in water or oil to preserve aerodynamic similarity, has flexible wings and is instrumented with tiny transducers to measure forces. For the first time, we can reproduce the flight of insects even smaller than fruit flies. The aerodynamics of such small insects is guaranteed to be laminar, which is deemed to be essential for validation. In the process of choosing a validation experiment, we showed that there is turbulent airflow over the wings of large insects and small birds. This invalidates the generally held belief that the functional dividing line between insects and birds is laminar versus turbulent flow. |
| Exploitation Route | Large-scale engineering simulations are crucial for virtual product development cycles, which are nowadays commonplace in most industries. Such integrated frameworks enable the design of better products by reducing the need for physical prototyping and keeping development times and costs low. Two of the current limitations in addressing industrial simulation needs are multiphysics computations, such as fluid-structure interaction, and solution of large three-dimensional problems. The implemented computational toolbox and the developed new techniques provide a fast and exceedingly robust framework for large-scale fluid-structure interaction problems. As such, the performed research will have an impact across diverse industries including bioengineering, aeronautics and environmental engineering. The developed computational toolbox named openFTL (open Finite Element Template Library) provides a validated framework for research and development in computational mechanics and engineering. Currently, it serves as the common software platform in our research team in Cambridge and beyond and forms the basis of six ongoing PhDs and one completed Masters project. These projects are in the broad area of fluid-structure interaction, high-performance computing, thermo-hydro-mechanical coupling and structural shape optimisation. We have two follow-on projects relating to insect flight and micro-air-vehicle design. The first is on the effect of flexibility on flapping flight efficiency; and, the second is on the development of variational multiscale based large-eddy simulation (LES) techniques. Both projects will use validation data from new experiments being conducted at Prof Ellington's lab. Additional mid-term drivers for our research are fluid-structure interaction in the context of subsea structures for energy generation and bioengineering. Of particular interest are the internal flows occurring in the human body and the related questions of optimal computational stent, graft and valve designs. |
| Sectors | Aerospace/ Defence and Marine,Chemicals,Education,Energy,Environment,Healthcare,Manufacturing/ including Industrial Biotechology |
| URL | http://www-g.eng.cam.ac.uk/csml/ |
| Description | This project enabled us to develop a computational toolbox and innovative techniques to simulate fluid-structure interaction problems. Our results are of interest to both engineering practice and science in general. In order to meet the simulation needs of our driving application insect flight and micro air vehicle design we developed several innovative techniques and implemented a scalable computational toolbox. Beyond our driving application, our research is relevant to a wide array of applications from engineering and science, such as medical hemodynamic assist devices and microfluidic membranes, just to name few. ----- The key novelty of our work is the use of immersed methods in which the fluid and solid meshes need not be compatible. We developed finite element based sharp interface tracking methods (using the Nitsche technique) which effectively eliminate the need for remeshing and lead to much more robust algorithms. The proposed subdivision-stabilised immersed b-spline discretization introduces a radically new approach for computing incompressible flows. Moreover, the typical loss of numerical stability in small cut-cells is circumvented by an original extrapolation technique. ----- The second novelty is the subdivision shell formulation for modelling insect and micro air vehicle wings, composed of veins/spars and membranes. Conventional shell finite element formulations are applicable either to thin or to thick shells and not ideal for simulating insect wing like structures. Based on our earlier experience with subdivision shells, we developed a new shell finite element, which is equally well suited for thick as well as thin shells. ----- The developed computational toolbox openFTL is beyond insect and micro air vehicle flight a general toolbox for research and development in computational mechanics. The core components of openFTL comprise over 160.000 lines of code. In addition, it relies on a number of open source third-party libraries, like Petsc or Metis. As demonstrated on HECToR (UK's high-end computing resource) openFTL is scalable up to thousands of processors. The development of openFTL has been leveraged by complementary funding from EU FP7 and Royal Society and Cambridge Trusts. ----- Finally, the tangible outcomes of our project include an experimentally validated large-scale simulation software for new research and industrial projects, publications in internationally leading journals and conferences and the training of about nine researchers. Moreover, the present project led to about seven invited seminar presentations at leading universities, which is perhaps a good indicator for the future impact. . Contribution Method: This project contributed to the UK and global research base in several key ways. First, the developed computational toolbox and techniques are of value for the design of micro air vehicles, the study of insect flight and beyond. Second, the unique combination of engineering and biological research enabled us to create an internationally unique forum in Cambridge for cross-disciplinary research on biological fluid-structure interaction. To this end, Professor Ellington's network was crucial in establishing us as recognised researchers in the insect flight and micro air vehicle communities. Third, from an educational viewpoint, this project produced several researchers and engineers, who are very competent in computational science and engineering. In addition to the two research associates and three PhD students we gave seven undergraduates and two school pupils the opportunity to contribute to the project in various ways. As a long lasting legacy, we hope that the success of the present project helped to raise the profile of advanced modeling and simulation on high-end computing platforms in engineering. Development and construction of the 3rd generation of flappers: ornithopter-like mechanisms that accurately follow the wing motions of insects while flow visualization and force measurements reveal underlying aerodynamic principles. The results are of academic interest to zoology and aerodynamics, and have practical application to the design of micro air vehicles. Our understanding of animal flight has been transformed by the use of flappers: ornithopter-like mechanisms that accurately follow quantified wing motions while flow visualization and force measurements reveal underlying aerodynamic principles. The first flapper, developed in our lab, simulated the wing motion of a large hawkmoth. It revealed a spiral leading-edge vortex that enhanced the lift. This was a novel and unexpected mechanism that has since been found for all insects that have been investigated, and it explains away the old chestnut that 'insects cannot fly.' Two second generation flappers were built by Fritz-Olaf Lehmann. These were flapped in oil to increase the aerodynamic forces while simulating Drosophila-sized insects. These studies greatly increased our understanding of the aerodynamics of flapping flight at very small size, but the design of the machines facilitated experiments only over a very narrow range of conditions. In this project we designed, built and tested a new 3rd generation flapper. Mechanical noise was reduced as far as possible by the use of jewel bearings, flexures rather than gears, and DC torque motors rather than stepping motors or commutated DC motors. Very sensitive force transducers with a good frequency response were designed to measure dynamic forces on the flapping wings to an accuracy of 0.1 g. The drive mechanism is so quiet that the transducer signals do not need any filtering or processing. Development of this machine was slower than we would have liked, and it has only recently been completed, but experiments have now started and we are confident that it meets all of our present and future requirements. It will support experiments that cover the size range of most insects. Given the advances in our knowledge of insect flight using flappers designed to study only large and small insects, a versatile machine that covers the whole range should have a large impact on the field. Beneficiaries: Academic research (animal flight and aerodynamics) and MAV technology Contribution Method: The work on insect flight has proved very useful for research and development of micro air vehicles (MAVs), a subject of significant commercial and military attention. The large conference and resulting book, 'Flying Insects and Robots', held during this project, showed that insects provide biological inspiration for current work on sensors, actuators, navigation and control systems, and aerodynamics. We believe that the new flapper studies will also have a considerable impact internationally on the MAV field. Finally, it could be said that some of us who study insect flight have necessarily become more expert in low Reynolds number aerodynamics than we might have anticipated. There had been rather limited professional interest in that Reynolds range before, and in order for our understanding of insect flight to progress we had to undertake basic aerodynamic research. Our appearance in the aerodynamic literature is perhaps our most wide-ranging contribution to the research base. We were the first lab to collaborate on a full 3-D unsteady CFD analysis of hovering insect flight, and our techniques and expertise are still well suited to the cutting-edge collaboration with Dr Cirak on fluid-structure interaction. |
| Sector | Aerospace/ Defence and Marine,Construction,Digital/Communication/Information Technologies (including Software),Education,Energy,Environment,Manufacturing/ including Industrial Biotechology,Security and Diplomacy,Transport |
| Impact Types | Cultural,Economic |
| Description | Fellowship for Yossef Elimelech |
| Amount | £19,000 (GBP) |
| Organisation | British Technion Society |
| Sector | Charity/Non Profit |
| Country | United Kingdom |
| Start | 09/2010 |
| End | 09/2011 |
| Description | HiFly: Direct numerical simulation of flapping flight on HECToR |
| Amount | £1 (GBP) |
| Funding ID | 8.000.000 computing core hours on Hector (UK's high-end computing resource) |
| Organisation | European Commission |
| Sector | Public |
| Country | European Union (EU) |
| Start | 11/2011 |
| End | 02/2013 |
| Description | Knowledge Transfer Secondment (KTS) scheme with Aquattera Energy Ltd |
| Amount | £24,000 (GBP) |
| Funding ID | EP/H500235/1 |
| Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
| Sector | Public |
| Country | United Kingdom |
| Start | 03/2012 |
| End | 10/2012 |
| Description | PhD Studentship for Matija Kecman |
| Amount | £58,000 (GBP) |
| Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
| Sector | Public |
| Country | United Kingdom |
| Start | 10/2010 |
| End | 10/2013 |
| Description | PhD Studentship for Musabbir Abdul-Majeed |
| Amount | £96,000 (GBP) |
| Organisation | Cambridge Commonwealth Trust |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 10/2012 |
| End | 09/2015 |
| Description | PhD Studentship for Musabbir Abdul-Majeed |
| Amount | £96,000 (GBP) |
| Organisation | Cambridge Commonwealth Trust |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 10/2012 |
| End | 10/2015 |
| Description | Collaboration with KAIST |
| Organisation | Korea Advanced Institute of Science and Technology (KAIST) |
| Country | Korea, Republic of |
| Sector | Academic/University |
| PI Contribution | Joint research on Stability and Controllability Improvements of Bio-inspired UAVs Using a Pseudo Flight Environment. We are using kinematic data from repeated bumblebee manoeuvres on a flight mill (the KAIST project) to drive the EPSRC flapper and measure the resulting forces. This is instead of the less reliable kinematic data from butterflies, as proposed originally. The high-speed camera from the EPSRC grant has been used heavily for the bumblebee project, and funds from the KAIST grant have been wired in return to support K. de Clerq in her 4th year. |
| Start Year | 2009 |
| Description | Collaboration with Prof Kenichi Soga in Cambridge |
| Organisation | University of Cambridge |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | The computational toolbox openFTL developed within this project is used for simulations of energy foundation piles and tunnels. In a joint project with Prof Kenichi Soga we are investigating large-scale thermo-hydro-mechanical (THM) coupled processes in the context of energy foundation piles and tunnels. |
| Collaborator Contribution | Prof Kenichi Soga is an expert in geotechnics and provided the know-how on energy foundation piles and tunnels. |
| Impact | The computations performed with the developed software form a large part of the PhD thesis submitted by Yi Rui. |
| Start Year | 2010 |
| Description | Collaboration with TU Delft |
| Organisation | Delft University of Technology (TU Delft) |
| Country | Netherlands |
| Sector | Academic/University |
| PI Contribution | K. de Clerq has remained in contact with the DelFly project at Delft University of Technology, where she did her MSc, advising new students on the project. |
| Collaborator Contribution | The DelFly is the most successful flapping wing micro air vehicle to date, and is an excellent subject for fluid-structure interaction. |
| Impact | Throughout the project this collaboration guided us in the choice of the computational problems which we studied. |
| Start Year | 2009 |
| Description | Collaboration with TU Munich |
| Organisation | Technical University of Munich |
| Country | Germany |
| Sector | Academic/University |
| PI Contribution | Collaboration on variational multiscale based large-eddy simulation (LES) techniques in fluid-structure interaction. We hosted in Cambridge Isabel Franck a Masters Student from Prof Wall's group at TU Munich and provided her with access to our large-scale simulation software. |
| Collaborator Contribution | The Institute for Computational Mechanics, headed by Prof Wolfgang Wall, at the Technical University Munich is one of the world's leading institutions in computational fluid-structure interaction. During her stay Isabel Franck has implemented in our simulation software a novel variational multi-scale large-eddy simulation technique specifically tailored for b-spline based immersed finite elements. |
| Impact | We have benefited from this collaboration in performing the simulations reported in our paper DOI: 10.1002/fld.3864. |
| Start Year | 2012 |
| Title | Computational toolbox - openFTL |
| Description | openFTL (open Finite Element Template Library) is a computational toolbox for research and development in computational mechanics. It consists of interdependent, complementary modules including, but not limited to fluids, shells and membranes and their coupling. openFTL's design closely follows the generic programming paradigm and is implemented using C++. It is scalable on multi-core, multi-processor computing platforms with several thousands of processors. The core components of openFTL comprise over 160.000 lines of code. In addition, openFTL relies on a number of open source third-party libraries, like Petsc, BDDCML, Metis and NLopt. The development of openFTL has been funded through the present EPSRC project and complementary EU FP7 and Royal Society projects and Cambridge Trusts PhD scholarships. |
| Type Of Technology | Software |
| Year Produced | 2014 |
| Open Source License? | Yes |
| Impact | In my research team openFTL serves as our joint development and simulation platform. Hence, all research in my team is based on openFTL. Beyond my research team, components of openFTL are also used by Prof Kenichi Soga (Cambridge), Prof Denis Zorin (New York University) and Dr Jakub Sistek (Czech Academy of Sciences). |
| Title | Development of a New Generation of Flappers |
| Description | A mechanical flapper has been developed for experimental validation of the predictions of fluid-structure interaction computations. Flappers are ornithopter-like devices that accurately follow the wing motions of insects while flow visualization and force measurements reveal underlying aerodynamic mechanisms. To create new data, we developed a 'flapper' that accurately mimics the wing motion of insects. It can fly either in water or oil to preserve aerodynamic similarity, has flexible wings and is instrumented with tiny transducers to measure forces. For the first time, we can reproduce the flight of insects even smaller than fruit flies. The aerodynamics of such small insects is guaranteed to be laminar, which is deemed to be essential for validation. In the process of choosing a validation experiment, we showed experimentally that there is turbulent airflow over the wings of large insects and small birds. This invalidates the generally held belief that the functional dividing line between insects and birds is laminar versus turbulent flow. |
| Type Of Technology | New/Improved Technique/Technology |
| Year Produced | 2012 |
| Impact | The flapper enables for the first time to measure the forces generated by insects even smaller than fruit flies. The generated data sets provide an invaluable input to future research in insect flight. |
| Description | Invited seminar speaker on fluid-structure interaction at seven UK institutions |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Other academic audiences (collaborators, peers etc.) |
| Results and Impact | Invited seminar presentations at Cardiff University, Durham University, University of Southampton (2x), University of Strathclyde, University of Cambridge, University of Oxford and BP Institute Cambridge. So far, Dr Cirak was invited by six UK institutions to specifically present the research outcomes produced in this project. The seminars were attended by faculty, postdocs, research students and undergraduate students. Increase in citations of our papers and requests for collaboration. |
| Year(s) Of Engagement Activity | 2011,2012,2013 |
| Description | STEM Outreach: Two Nuffield Placements |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | Yes |
| Geographic Reach | Regional |
| Primary Audience | Schools |
| Results and Impact | During the summer 2012, we hosted two pupils from local schools, Fergus Waugh and Nick Tikhonov, in my lab. Nick is a pupil at the Parkside school (state) and Trys is a pupil at the Perse school (independent), both in Cambridge. Nick is now studying Computer Science at St Andrews. |
| Year(s) Of Engagement Activity | 2012 |
| URL | http://nicktikhonov.tumblr.com/. |
| Description | Six MEng projects on aspects of flapping flight |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Undergraduate students |
| Results and Impact | We offered six MEng "fun" projects of educational nature on different aspects of flapping flight. The six offered projects are: - Structural performance of insect wings, Thomas Greenwood (2008/09) - Physical simulation and animation on GPU, Matija Kecman (2009/10) - Development of a biomimetic micro-air-vehicle, Trys Negus (2009/10) - Computational analysis of flapping micro-air-vehicles, Thibault Flinois (2010/11) - Resonant structures for flapping micro-air-vehicles, Jiho Han (2010/11) - Resonant structures for flapping micro-air-vehicles, Jiho Han (2012/13) Two of the students, Matija Kecman and Clare Hall, started a PhD in Cambridge after finishing their projects. |
| Year(s) Of Engagement Activity | 2008,2009,2010,2011 |