Read Aerogram Newsletter text version

Aerogram

Bringing news from the horizon

June 2006 Volume 11 Issue 6

Aerospace in the news

Dr Paul Marshall, Head of Cranfield University Aerospace

© AIRBUS S.A.S. 2005 Photo by F Espinasse

Shortly after the publication of the last issue of Aerogram the first twin-deck Airbus A380, the world's largest commercial airliner, was to be seen making several demonstration flights at the Paris Air Show. Today, as I write this introduction, an A380 aircraft is at Heathrow Airport, part of the comprehensive test programme which continues throughout this year with first customer deliveries, starting with Singapore Airlines, scheduled for late 2006. Meanwhile the other leading commercial airline manufacturer, Boeing, has been busy developing its own new product, the 787 `Dreamliner'. This new mid-sized airliner is to exploit composite materials to a level unseen thus far in commercial airliners. Boeing has designed the 787 to satisfy the point to point market whereas the Airbus 380 is designed for hub to hub operation. More recently Airbus is countering the 787 with a new offering, the A350, the design for which should be finalised later this year. These new aircraft projects are embracing not only many new technological developments but also massive changes in business models including globalisation of manufacturing base and risk sharing with the supply chain. The recovery from the dip in air travel following the events of 9/11 appears to be complete, with huge projected growth despite record fuel prices. In the world of business aviation we can look forward to the imminent introduction of a number of new aircraft that have come collectively to be called VLJs (very light jets). The Adam A700, Cessna Mustang, Eclipse 500, Embraer Phenom 100 and the HondaJet to name but a few. This is a very healthy market helped in part by developments in new operational models including `fractional ownership'. The healthy order book for large and small commercial aircraft alike is indicative of the projected growth in passenger numbers. These projections are good news for the aircraft manufacturers but bring increasing concerns over local and global environmental issues, congestion, safety and security. In the Defence arena, development of the A400M transport aircraft continues. The aircraft, expected to fly in 2008 and enter service in 2009, will incorporate many technological advances including composite main wing spar and skins and four new Europrop International TP400-D6 engines, the most powerful turboprop powerplant in the Western World. The first production line F-35 Joint Strike Fighter was rolled out early in 2006. Although the contentious issue of technology access still plagues relationships between the UK and the US, the milestones for this exciting project appear to be on schedule. There continues to be a massive growth in activities related to unmanned air vehicles (UAVs). Roles include tactical reconnaissance, targeting and battle damage assessment,

continued overleaf

Á

www.cranfield.ac.uk/cua

June 2006 Volume 11 Issue 6 www.cranfield.ac.uk/cua

Aerospace in the news

electronic warfare and payload delivery. Operational use by the military is now in the public domain. There is also continued growth in commercial applications in agricultural crop spraying, electrical power line checking, atmospheric research, data-link relay and traffic/security surveillance. To add to the plethora of platforms we are all aware of ­ Global Hawk, Predator, Hermes and Phoenix ­ earlier this year we learned of several more as BAE Systems revealed the extent of its `black' programmes on UAV technology. Programmes like Raven and Herti can be traced back to 2001, but information on these has only recently come into the public domain. Development of UAV technology seems relentless with much emphasis on autonomy and operation within the Network Enabled environment. Since the last issue of Aerogram the Space Shuttle fleet has re-entered service following the destruction of Columbia in 2003. The International Space Station continues to be inhabited, serviced mainly by Russian Soyuz spacecraft. At the other extreme of manned spaceflight, space tourism is rapidly becoming a reality with development of the Ansari X prize-winning concept by Virgin Galactic. In planetary exploration ESA's Venus Express entered orbit earlier this year, Mars Express continues to perform despite the demise of the Beagle 2 lander. NASA's Mars Rovers continue to perform well, exceeded their design life many times over and, in January 2006, NASA launched the New Horizon probe, the first probe to be sent to explore Pluto.

(continued)

Contents

Cranfield University has been involved, directly or indirectly, in many of these programmes; there has also been significant activity in the operational, environmental and socioeconomic issues arising from these recent developments in the aerospace and aviation industrial sector. You will see some of the highlights in the articles that follow. A major part of this edition of Aerogram is devoted to the application of Computational Fluid Dynamics; while in one respect this highlights the depth of expertise in one very focused discipline, you will also see the breadth in terms of the areas of application. Finally, just a few words about changes within Cranfield University since the last edition. The contract confirming Cranfield as the Academic Provider to the Defence Academy at Shrivenham was signed. Natural Resources activity at Silsoe is being transferred to the Cranfield campus as a department within the School of Industrial and Manufacturing Sciences ­ itself now renamed the School of Applied Sciences ­ with health-related activities being brought together to form a new school, Cranfield Health, based on the Cranfield campus. Continued investment in infrastructure and facilities has resulted in the opening of the £2.5m Lightweight Structures laboratories comprising composites manufacture, polymer and ceramics processing and state of the art impact and crash testing facilities. And, in February this year, the university was awarded The Queen's Anniversary Prize for Higher and Further Education for its Fellowship in Manufacturing.

4

Cranfield news

8

Computational fluid dynamics aerodynamics and micro/nanotechnology

14

Gas turbine engine nacelle ventilation and fire events

33

More news

17

Computational rotor aerodynamics at Cranfield University

24

Integrating CFD and experiments in aerodynamics

29

CFD in aerospace design

35

Short courses

The Kestrel shows the way

Award for `impeccable record'

The delighted Cranfield Airport team scooped the Aircraft Owners and Pilots Association's (AOPA) award for its huge contribution to general aviation at all levels. The awards committee noted that `although the airport is often busy, visitors flying in are made welcome, and Cranfield's record in accepting genuine weather diversions with calmness and efficiency is impeccable'.

Scale-model prototypes completed for Boeing

Working virtually in secret for four years, Cranfield Aerospace has completed two scale-model prototypes of the Blended Wing Body (BWB) aircraft for Boeing. Constructed of advanced carbon fibre composites, foam and honeycomb and with a 21ft wingspan, they will enable Boeing to explore the flight dynamics, control laws and aerodynamic issues of the radical new design.

Pictured is Senior Lecturer Dr Howard Smith with the Kestrel sub-scale flying demonstrator, used to partially validate the Blended Wing Body concept. The Kestrel weighs some 130kg, has a wing-span of about 5m and is powered by two small AMT Olympus jet engines. It was designed and partially completed by part-time students of the Cranfield MSc in Aircraft Engineering, many of whom are BAE Systems employees, and was completed by their company. That it was then successfully flown by BAE Systems in March 2003 is something only recently made public.

Wind-tunnel tests on the first vehicle at the NASA Langley Research Centre have just been completed, and the data will now be used to examine how the multiple surfaces on the wing can best be used in manoeuvring the aircraft. Test flights will then begin later in the year at the NASA Dryden Flight Test Centre in California. Weighing some 400lb, each prototype is powered by three JetCat P200 turbojet engines capable of 220N (49lb) thrust and will be capable of flying at up to 120knots and an altitude of 10,000ft. With potential as a long-range bomber or a flight refuelling tanker, the full-scale version could be capable of flying non-stop around the world at 600knots.

(l-r: back) Bob Pooley (Chairman Pooleys Flight Equipment), Mandy Hughes (Airport Admin), George Done (Chairman (AOPA), David Wilkins (Airport Director), Clifford Friend (Deputy V-C), Russ Edwards (Serco), Lindsay Marchant (Leading Fire Fighter), Ken Lampard (Sub Officer). (l-r: front) Sue Amor (Airport Accounts), Janet Phillips (Airport Accounts), Susanne Hastings (Fire Fighter)

Picture courtesy of Boeing

03Aerogram Bringing News from the Horizon

June 2006 Volume 11 Issue 6

3

Cranfield news

Cranfield news

New Shrivenham contract with MOD signed

The Defence College of Management and Technology (DCMT) was created in November 2004 as a component college of the Defence Academy, replacing the Royal Military College of Science (RMCS). And, following the announcement late last year of the signing of the 22-year Academic Provider contract for postgraduate education at the Defence Academy, Cranfield University at Shrivenham was similarly renamed the Defence College of Management and Technology. DCMT is a military and academic community which builds on the foundations of the knowledge, expertise and experience of RMCS, but with a broadened scope ­ providing education in management, strategic leadership, technology and related aspects of security studies for the three armed services and the Ministry of Defence civil service. The courses will offer a mix of residential and blended learning programmes to suit the needs of today's students. Plus, in

Researching the effect of air flows within bomb bays

Research Officer Simon Ritchie has been sponsored by missile manufacturer MBDA UK to investigate the effect of air flows within aircraft bomb bays ­ or cavities ­ for use in future `launch prediction software'. In recent years, modern fighter jets have been designed to carry their weapons in bays inside the aircraft, rather than on their wings. Not only does this increase the aircraft's performance, it also improves its body shape, reducing its radar signature ­ its `footprint in the sky' ­ and making it more difficult to pick up on enemy radar and, therefore, less of a target. This change in design, however, brought problems concerning the size of the bays and the air flows within them once the doors have been opened to release the weapons ­ issues that MBDA UK has sponsored Cranfield to look at. To open its doors and release its bombs, a fighter jet needs to slow down from supersonic speeds to around 85% of the speed of sound. In deep bomb bays, air passing over the bay can cause vibrations as the doors are opened. "Rather like blowing over a bottle top," explained Simon. This vibration not only increases an aircraft's signature and vulnerability, but can also damage sensitive electronic systems housed within the weapons bay. Worse still, some aircraft house their weapons in shallow bomb bays. As the jet opens its doors and releases its weapons, it has been known for pressure differences within the bay to make the bombs point upwards and career into the underside of the aircraft on release. Cranfield research has centred on taking laser-based measurements of air flow within these bays, using scale models and a high-speed wind tunnel. Involving taking measurements of air flow speeds similar to those experienced when a jet opens its door, this is the first university to do so within a whole bay in a single measurement.

recognition of the increased level of management and leadership programmes required under the new contract, Cranfield School of Management will be a major contributor to the course delivery. To address the increasing emphasis of distance learning within the course programmes, Cranfield will work with the Open University as its principal subcontractor. The world-renowned King's College, London; academic provider to the Joint Services Command and Staff College at Shrivenham, will also be providing courses in defence studies as part of the contract. Professor Frank Hartley said: "We look forward to continuing our successful partnership with the armed forces and the wider MOD. The awarding of this contract shows that, despite stiff opposition from other providers, Cranfield remains at the forefront of management and technology education."

Combining the aerospace technology of our applied science and engineering schools

Cranfield has combined the aerospace technology expertise of its applied science and engineering schools to launch the first academically recognised qualification in airworthiness. Professor John Bristow, a former Senior Manager with the UK Civil Aviation Authority and a Cranfield Visiting Professor, said: "In today's competitive market, airlines need to strike a better balance between safety and cost in order to maximise performance. To achieve this, a detailed knowledge of airworthiness issues is needed by all engineers working in the field of aviation safety, whether in the design and production of aircraft or their maintenance, modification and repair." Run on a part-time basis, the course provides a fundamental understanding of airworthiness related to the wide spectrum of technologies met in aerospace. Steve Swift of the Australian Civil Aviation Safety Authority said: "I welcome Cranfield University's new Airworthiness MSc, especially its regulatory emphasis. The course will, no doubt, benefit from the practical experience gained from the university's long and strong associations with manufacturers, operators and authorities, both civil and military." Course director Dr Ken Ramsden said: "Cranfield academics, together with experienced regulators and industry experts, are combining their expertise to explain the safety aspects of current airworthiness regulations in relation to the background technologies."

F22 displaying weapons bays (above) F22 Raptor (right)

4

Aerogram Bringing News from the Horizon

June 2006 Volume 11 Issue 6

5

Cranfield news

Cranfield news

Cranfield Airport

Cranfield University is unique among UK universities in having its own operational airport as an integral part of our practically orientated postgraduate aerospace programmes. The asset is of strategic importance both in attracting students from global markets to our aerospace programmes and in enabling the university to undertake major projects, such as converting a Harrier jet to fly-by-wire. Two indications of the university's continuing commitment to aerospace programmes are the current £4m investment in the refurbishment of Hangar 3 and the recent acquisition of a Jetstream 31 (pictured) to replace the old Jetstream MK1 as a flying laboratory. Infrastructure and operational costs at all airports are high, however, and the university could not sustain an airport solely for its own use. Other operations, training, business aircraft and so on are, therefore, encouraged to keep the airport viable. Now, to further encourage the already increasing numbers of high added-value business aircraft, the university, together with its development partner St Modwen, has committed a significant area of land for an Air Park. The Air Park will comprise bespoke hangarage and other accommodation to occupiers' requirements, together with a management office dealing with the dayto-day logistics, maintenance and security. It will be conveniently located at the northern end of the Airport, easily accessible from both the taxi-way and College Road. While the university offers similar airport terms and conditions to other UK airports and can make no guarantees about the future, such a significant investment demonstrates its confidence in the airport operation for the foreseeable future.

A most compelling study

There are considerable economic as well as environmental benefits associated with regional air travel. This was apparent when Dr Peter Morrell presented the findings of a Cranfield University study at Boeing's 787 Dreamliner Community Awareness Roadshow at Luton in March. The 787 is a key element of Boeing's strategy, and one of the key aims of the community awareness programme is to reinforce the idea that people want to go where and when they want to go...point to point, not through hubs. Entitled `The environmental cost implication of hub-hub versus hub-bypass flight networks', this compelling study set out to answer the question of how the balance of hub-to-hub versus bypass would change if the airlines were required to

account for all the noise and emissions costs they incurred. It outlined the fact that the noise and emissions social cost impact of the pointto-point networks was significantly lower than the hub-to-hub in all cases. The message that point-to-point is not only less expensive to operate but is also better for the environment was particularly well received by Luton's airport community. With an attendance of more than 60, including the constituency MP for Luton North, Kelvin Hopkins, and Margaret Moran, MP for Luton South, the group of stakeholders also included local area councillors, industry, the airport operator, trade associations and, of course, customers.

The event followed on from a wellreceived London roadshow in 2004 which demonstrated the environmental, economic and passenger benefits of the 787. A significant component of the 787 story as it relates to the UK has been the industrial and scientific contribution made by British companies and organisations. Throughout the last two years, Boeing has worked with its partners and suppliers on the 787 programme to illustrate the longterm, high-value jobs that are being supported through the UK's contribution to the 787's development. Wherever geographically possible, partner companies have delivered a segment during the presentation and, at Luton, QinetiQ described a range of capabilities that are being applied in the acoustics and aerodynamics arenas.

How resilient is the air transport industry? The Airport Planning and Management MSc comes to Cranfield University

Following an approach by Loughborough University, Cranfield was delighted to take over the running of its successful Airport Planning and Management MSc. The only masters course in airport planning to be taught in English, it complements Cranfield's existing programmes in Air Transport Management and Safety and Accident Investigation and is aimed at people seeking a career in airport planning and the management of airports. Aviation economist Dr Romano Pagliari, lecturer on airport economics and business issues in the Department of Air Transport is behind the new venture. Dr Pagliari provides research support to Cranfield's consultancy activities, recruits for the full-time MSc in Air Transport Management, and organises short courses on airport economics and airport commercial revenue development. He is also co-organiser of the Forum on Air Transport in Remoter Regions. Another new course for this year is Avionic Systems Design, an option of the Aerospace Vehicle Design MSc. A two-day conference organised by the university and the RAeS Air Transport Group set out to discover how resilient the aviation community really is. Professor Trevor Taylor of our Resilience Centre opened proceedings with a definition of resilience as "an ability to deal with shock, minimise the impact, and recover rapidly." He also noted the long-standing attractiveness of aviation to terrorists. Other speakers, including keynote speaker Bruce George MP picked up on , Professor Taylor's overall theme of the vulnerability of air transport to terrorism in all its manifestations, warned against focusing too exclusively on a single threat such as Islamic fundamentalism, and looked at the threat of chemical, biological, radiological and nuclear attacks from a variety of single interest terror groups. The themes of `how resilient is the public?' and the kind of society we are defending resurfaced in later the presentations, while talks by airline security representatives emphasised the co-operation ­ sometimes with deadly business rivals ­ that goes on in countering terrorist threats. The relevance of the new media and the effect of terrorism on business were the topics of other talks. Delegates concluded that the global aerospace industry today is both less and more resilient ­ less in that news travels faster and can be exaggerated; more as a result of the long-term practical measures.

Picture courtesy of Airbus

Finally, it was stressed that all discussion of resilience and security must deal with the concept of risk, and that the key to resilience may be adaptability, self-reliance and individual responsibility ­ allowing rapid rebuilding when the inevitable happens.

June 2006 Volume 11 Issue 6

6

Aerogram Bringing News from the Horizon

7

Computational fluid dynamics, aerodynamics and micro/nanotechnology

Professor Dimitris Drikakis FRAeS, FIoN, Head of the Aerospace Sciences Department and Head of Fluid Mechanics and Computational Science Group (FMaCS), School of Engineering

Figure 2: Computational aerodynamics of intake and exhaust geometries of helicopter engine (Quaranta & Drikakis, 2005). The results have been obtained in the framework of the EU project Friendcopter. The upper plot shows a schematic of the intake geometry and pressure distributions near the engine, while the lower plot shows a schematic of the ejector and temperature distributions as obtained by the simulations.

Comprehensive reviews undertaken by the aerospace sector have identified a number of challenges that must be addressed in the next 50 years. The issues include noise, emissions/air quality and operational impact on climate change; eg reduced drag for lower fuel burn per passenger mile. If these are to be successfully dealt with in future aircraft and engine designs, significant technological barriers must be overcome in order to produce appropriate airframe and propulsion system performance at acceptable cost levels. To address the above issues radical improvements in our level of understanding and design capability in technical disciplines such as aeroacoustics, laminar flow control, shear layer manipulation and aerodynamic performance of advanced wings and blended wing/bodies, are required. Further, a much more detailed grasp of the instabilities and dynamical behaviour of these internal and external flow systems are demanded than has hitherto been possible. In the last three decades, the development of computers and associated computational sciences has been staggeringly quick. The advent of parallel supercomputers as being distinct from the supercomputers of yesterday presents a continual challenge to the working scientist and engineer and has provided them with phenomenal computing power at their fingertips. The supercomputer of a decade ago now rests comfortably in the lap of the traveller sitting on an airplane or at the corner coffee shop. With this power, the advances in numerical methods allow a tremendous degree of simulation capability to be easily accessed. In this paper, we summarise research activities in the fields of computational fluid dynamics, computational aerodynamics/gas dynamics, as well as computational micro and nanotechnology, which are conducted by the Fluid Mechanics and Computational Science (FMaCS) Group at the Department of Aerospace Sciences. FMaCS is currently engaged in a large portfolio of research activities in fluid mechanics modelling and simulation including: · advanced wing technologies · helicopter aerodynamics · rotary wing · intake and exhaust geometries · flow control · aeroacoustics · fundamental research on transition, turbulence and instabilities · heat transfer · shock waves and associated instabilities · multi-phase and multi-component flows · flow, mass transport, heat transfer and chemical reaction processes in micro and nanotechnology

· processes and technologies in biology and medicine · microfluidics · membrane technology · environmental flows · development of computational science methods for scales ranging from continuum to atomic · numerical optimisation · high-performance parallel computing. Some of the above research activity is presented below. The FMaCS group is engaged in the development and implementation of Reynolds-Averaged Navier-Stokes (RANS), Large Eddy Simulation (LES) and Direct Numerical Simulation (DNS) modelling techniques. Particular emphasis is presently given on the LES techniques which seem to provide the best compromise in terms of accuracy and computational cost. The use of DNS for studying unsteady flows, which are of interest in aeronautical and aerospace applications, at high Reynolds numbers (105-107) is well beyond foreseeable computing power. For example, to compute the flow around an aircraft for one second of flight time, using a supercomputer of 1012 Flops, it requires several thousand years and 1016 grid points. On the other hand, RANS models encompass a number of assumptions, which hinder understanding of complex flow physics especially in time-dependent flows. Turbulence and transition, and its impact on aerospace aerodynamics, remains a fundamental area where continued research is vital if industrial aerospace Design Challenges are to be tackled successfully. FMaCS participates in the MSTTAR DARP (Modelling and Simulation of Turbulence and Transition for Aerospace, Defence Aerospace Research Partnership) programme, which was established in March 99 to promote research on the understanding and

modelling of turbulent flow phenomena of importance for industrial aerospace applications. The MSTTAR DARP is a collaboration between industrial organisations (BAE Systems, Rolls-Royce, QinetiQ), MOD, DTI, EPSRC and six academic institutions. FMaCS participates in the Configuration Design challenge. Future military configuration and UCAV (unmanned combat air vehicle) studies, give rise to wing shapes with unconventional leading and trailing edge angles of sweep as well as complex wing-body blendings. The accurate modelling of such flows is essential if any computational approach is required in design activity aimed at manipulating or controlling flow over configurations, with the goal of removing control surface dependency without penalising performance, or for low observability (LO) assessment. The flow regimes of importance are (i) highly skewed or separated boundary layers and (ii) leading edge separation resulting in shear layer roll-up. The flow over a highly swept wing at incidence is characterised by separation from the leading edge, which may be of the closed (bubble) or open (vortex) type, depending upon the sweep angle. At present, there are no theoretical models capable of predicting such separation with any degree of certainty, nor can the nature of these changes in the flow behaviour be convincingly explained. FMaCS currently uses with success LES in conjunction with high-resolution methods to model and simulate highly swept wing flows at both high and low Reynolds numbers. Using these techniques, we aim to improve physical

Figure 1: Computational results for flow around a highly-swept wing at a=14° angle of incidence and Reynolds number of 250,000 (based on the mean chord length) using LES in conjunction with high-resolution methods. In the second line the computational results (Hahn & Drikakis, 2005) are compared with the experimental oil visualisations provided by Shanying Zhang & John Turner at Manchester University.

understanding of the flow structures and their changes and formulate a quantitative model based on the global balance of vorticity, explaining changes between the two flow regimes. Furthermore, we aim to provide industry with a new family of CFD (computational fluid dynamics) techniques, based on high-resolution methods, which will be able to simulate complex flow physics (incl. transition and turbulence) around and inside aerodynamic geometries at a fraction of the computing time compared to the current state-of-the-art. Indicative results from the above research are shown in Figure 1, where the flow around a highly swept wing designed for the purposes of the MSTTAR project has been simulated using LES and highresolution methods at a moderately coarse grid (about 3M grid points) and the results have been compared with the experimental oil visualisations provided by Manchester University. The results reveal that using LES based on high-resolution methods, the basic flow physics including the structure/path of the vortex, transition line and trailing edge separation, are captured similar to the experiment. Computational aerodynamics research is also carried out by the group in relation to several other internal and external flows in the framework of EU and EPSRC funded research programmes. The research aims to improve/optimise aerodynamic designs in terms of overall performance. Figure 2 shows results from computational aerodynamics simulations of intake and exhaust geometries of helicopter engines, while Figure 3 shows results from simulations of aerofoil design with smart/active trailing edge.

8

Aerogram Bringing News from the Horizon

June 2006 Volume 11 Issue 6

9

Computational fluid dynamics, aerodynamics and micro/nanotechnology (continued)

Trailing edge geometry

Modelling and simulation work is also conducted on applications pertinent to micro/nano-technology and nanoscience. The field of nano-technology goes back to Richard Feynman's ideas on: "There's Plenty of Room at the Bottom." Often, nanotechnology is concerned with problems at dimensions of 0.1 to 100nm; however, the boundary between nano and micro technology becomes less distinguishable in terms of physical scales as the dimensions increase. Examples of nanotechnology include design of new materials with applications from medicine to mechanical, chemical and aerospace engineering; chemical and bio-detectors; new generation of lasers; systems on a chip; carbon nano-tube products; nanoparticle reinforced materials; thermal barrier; ink jet systems; information recording layers; molecular sieves; high hardness cutting tools and fabrication.

Figure 5: Mach 1.2 shock wave passing through a block of dense sulphur hexaflouride producing Kevin-Helmholtz and Richtmyer-Meshkov instabilities. The initial condition (top left corner, t = 0) shows right facing shock and the block of sulphur hexaflouride; time increases from left to right and top to bottom as follows t = 0.5ms, 1.2ms, 3.9ms, 5.2ms. The last image is a sample at t = 1.4ms for a higher-resolution simulation. The computations have been performed using LES in conjunction with highorder/high-resolution methods (Thornber & Drikakis, 2005).

OA312P5 OA312 OA312M5 Figure 3: Computational aerodynamics of aerofoil designs with smart/active trailing edge (Zachariadis & Drikakis, 2005). The results have been obtained in the framework of the EU project Friendcopter. The plots show the effects of the trailing edge device on the shock formation and shock/boundary-layer interaction around the aerofoil.

OA312P5

OA312

OA312M5

Another area of research in which the FMaCS group is currently engaged is the area of computational aeroacoustics (CAA) both for aircraft and helicopter configurations. In the framework of the EU project Friendcopter, the group develops computer models based on acoustic analogy approaches (Ffowcs-WilliamsHawkings). Furthermore, the group plans to couple these approaches with advanced CFD methods such as LES as well as with boundary element methods for studying aeroacoustic problems both for internal and external configurations. An example from the implementation of these techniques is shown in Figure 3 relating to the prediction and noise abatement procedure for helicopters. In order to allow helicopters to fly close to urban areas during the day or night it is necessary to implement noise abatement procedures and design characteristics. One of the major sources of noise in a helicopter is the rotor blades. Figure 4 shows results from the application of the Cranfield aero-acoustics code for predicting helicopter rotor noise. Furthermore, the group is conducting research on shock waves, turbulence and turbulent mixing as well as associated instabilities such as Richtmyer-Meshkov and Rayleigh Taylor instabilities. Research in this area is supported by AWE, EPSRC and MOD. The work aims to investigate complex flow physics, turbulence and turbulent mixing in particular occurring when shock waves interact with material interfaces, as well as to provide a deeper understanding of the intrinsic physical properties encompassed by modern highresolution/high-order CFD methods. It is

motivated by a broad area of applications in science and engineering, including inertial confinement fusion (ICF), nuclear devices, detonation, supersonic combustion (scramjets), counter-gradient transport, instability of collapsing gas bubbles, atmospheric temperature inversions as well as problems in astrophysics (mixing of fluid in a supernova). Figure 5 shows simulation of a shock wave at Mach 1.2 passing through the interface between a dense and light gas (density ratio 5:1). The interface is perturbed with a twodimensional wave to investigate the growth of the mixing layer as a function of the size of the initial perturbation. As the shock passes through the interface instability is formed which distorts the surface into a characteristic mushroom shaped bubble with a ring vortex at the head. EPSRC-funded computational research is also performed in the field of microfluidics and membrane technology. This work complements experimental research that is carried out at IRC of Queen Mary, University of London. The research aims to develop advanced computational models for microfluidics applications and to investigate the effects of diffusion and mass transport occurring in microfluidics in relation to membrane formation across the interface of different fluids. The objective is to use multiphase laminar flow patterning as a method of microfabrication inside microchannels. Microchip or microreaction technology has been focusing much attention in recent years. Various kinds of effective chemical processes have been successfully integrated onto microchips to realise micro total analysis systems (µTAS) or microreactors (lab-ona-chip). Controlled laminar flows in micro-

fluidic devices are currently utilised to great benefit in a number of industrial applications including diffusion-based separation and detection, solvent extraction, mixing and hydrodynamic focusing. Figure 6 shows computational results from aqueous-organic interfaces for in situ formation of polymer membranes and investigations of reagents diffusion in microchannels aiming at problemdependent optimisation of microfluidic cell designs.

Figure 4: Aeroacoustics modelling of rotorcraft based on Ffowcs-Williams-Hawkings approach and the HelicA CAA code (Loiodice, Drikakis, Kokkalis 2005). The plots from left to right and top to bottom show the directionality of the noise propagation as generated by the helicopter rotor; discrete model of the blade and the microphones grid on the ground; contours of sound pressure generated from the blade; sound pressure history received by one of the microphones of the grid. The results have been obtained in the framework of the EU project Friendcopter.

Figure 6: The top plots show results from computations of aqueous-organic interfaces for in situ formation of polymer membranes using different microfluidics design cells. The bottom plot shows diffusion of reagents in microchannels (Shapiro & Drikakis, 2005).

Computational modelling and simulation of transport processes on nano scales are important because they can provide a more detailed picture of the transport processes than laboratory experiments and, additionally, they may be the only means to advance nanotechnology applications because of the difficulty to perform experiments on nanoscales. In respect to simulations, one of the challenges is that the computational approach may vary depending on the physical scales involved in the problem. For scales below 10-10m (ie 0.1nm=1Å), a quantum mechanical description is required. Quantum mechanical calculations are, however, complicated, and thus this theory is not broadly used. For scales larger than 0.1nm, the macroscopic behaviour of liquids can be calculated by the molecular dynamics (MD) approach. MD simulations involve scales of 100Å, in space, and of 10-9s, in time. The MD approach has been used in a variety of applications including liquid flows, non-equilibrium shock waves, polymers, proteins, shock dynamics and crack propagation, membranes, lubrication, wetting, and coating problems. The range of length scales of 10nm (or above) and time scales of few nanoseconds (or above) is a grey area for all computational approaches. The challenge is not only to provide an accurate solution to the problem in question but also to obtain this solution in the most cost effective manner. MD simulations are primarily used in

June 2006 Volume 11 Issue 6

10

Aerogram Bringing News from the Horizon

11

Computational fluid dynamics, aerodynamics and micro/nanotechnology (continued)

limited to time scales on the order of 10ns, using a desktop workstation. In the FMaCS we develop multiscale modelling approaches that combine different computational strategies to simulate the dynamics of materials as well as flow and mass transport processes from nano through micro to macro scales (Figure 7). The multiscale models and simulation strategies developed by the group are currently used in a number of practical applications spanning from material science to flow control and biomedical applications. Indicative examples from these applications are provided below. Figure 8 shows simulation of flow control devices on micro/nano scales using molecular dynamics. The aim of this investigation is to understand the phenomena occurring on micro/nano scales and provide information for optimising the fabrication of flow control devices on such scales. Figure 9 shows simulations for another problem, which concerns molecular cluster-cluster collisions. The study of atomic and molecular clusters is an active area of research in the field of nanosciences, both in the theoretical and experimental context. Nanoclusters offer a broad range of interesting possibilities not only in relation to material sciences, where they can be used to construct coating surfaces or new materials, but also in environmental and climate science problems, such as homogeneous gasliquid nucleation resulting in molecular clusters of supersaturated vapour, as well as in chemistry because of their high reactivity. In Figure 9 the large cluster has a diameter of 50Å and the small one of 14.5Å, comprising 10,973 and 309 atoms, respectively. Another recent application where the FMaCS group has been involved is the absorption of gases on carbon nanotubes. Even though carbon nanotubes have shown several outstanding physical properties, particularly concerning gas adsorption, a production technique is not available, which restrains experimental studies of these devices. Numerical simulation is necessary, but quantum mechanics, which is mainly used to study gas adsorption, is computationally expensive. In the last few years the group has developed molecular models for carbon nanotubes in order to study the influence of the structure and diameter of

12

Aerogram Bringing News from the Horizon

the nanotube on adsorption energy and equilibrium distance. Furthermore, we have conducted studies for the adsorption energy with respect to the amount of adsorbed gas on a single nanotube; differences of the adsorption process on an open-ended nanotube and on an infinite nanotube, as well as adsorption on nanotube bundle. Figure 10 shows results from molecular dynamics simulations (Labois, Kalweit & Drikakis, 2005) for the adsorption of argon on the walls of a carbon nanotube bundle. Finally, the group is engaged in the development and implementation of computational nanoscience models in medicine and biology. Research is conducted in connection with shockwave chemotherapy, drug targeting, blood perfusion, research pertinent to anaesthesiology as well as contribution to medical diagnosis and design of medical equipments. Research in the above areas is supported by NHS Hospitals and companies specialising in medical equipments. An example from our recent research on extracorporeal shockwave chemotherapy is shown in Figure 11. In this figure we show results for the interaction of a shock wave with a biological membrane (lipid) membrane aiming at understanding the biological and physical mechanisms associated with the reduction of cancer proliferation during the implementation of shock wave chemotherapy. Most of the computational models and applications presented here have been implemented in the Cambridge-Cranfield High-Performance Computing facility. The potential of advanced computational methods and models for scientific and engineering applications is vast and the results presented in this report demonstrate only a fraction of the actual capabilities available within FMaCS at the Department of Aerospace Sciences, School of Engineering.

Figure 7: Schematic of multiscale modelling and simulation strategy undertaken by the Fluid Mechanics & Computational Science Group in connection with the computational nanotechnology activity. The computational strategy models the material, micro/nano flow and mass transport as well as macroscale continuum processes including chemical reactions.

Figure 8: Micro/nanoscale simulation of pulsating flow through a cavity orifice. The investigation is pertinent to flow control devices on micro and nano scales (Kalweit & Drikakis, 2005).

situations in which the continuum approach (or the statistical methods such as Monte Carlo or Boltzmann equations) is inadequate to compute important flow quantities. For example, one could use classical MD to simulate flows in channels with dimensions of 1µm, but at an enormous computational cost. Moreover, the larger the sizes of the geometry, the more molecules need to be used, thus increasing the computational cost. Because of the above constraints, simulations even of simple liquids at length scales on the order of 15nm are

Figure 9: Evolution of the cluster-cluster collision at an impact velocity of 2000m/s. All figures show a central slice of the cluster with a thickness of 20Å at different time instants (Kalweit & Drikakis, 2004).

Figure 10: The left plot shows a schematic of the adsorption sites on a carbon nanotube bundle, while the right plot shows results from molecular dynamics simulations (Labois, Kalweit & Drikakis, 2005) for the adsorption of argon.

Figure 11: Molecular dynamics simulations of the interaction of a shock wave induced by an ultrasound device with a lipid membrane. The plot shows the change of the lipid mass centre in time and this result is used in analysing diffusion through the membrane (Lechuga & Drikakis, 2005).

June 2006 Volume 11 Issue 6

13

Gas turbine engine nacelle ventilation and fire events: experimental and numerical simulation

Dr Phil Rubini, Department of Aerospace Sciences, School of Engineering

EXHAUST

GEARBOX

The fire safe design of accessory zone layout and ventilation in aircraft engines has traditionally been founded on a combination of past experience and standardised fire testing of individual components. Through design necessity the annular region between the outer wall of the engine nacelle and the fan casing on a modern high by-pass ratio gas turbine engine is often highly congested. This volume will typically contain a range of engine accessories and equipment, together with systems handling the cooling and supply of lubricating oil and fuel. The combination of rotating machinery, for example, in the gearbox and starter systems, with electrical equipment and hydrocarbon fuels therefore introduces a potential fire hazard into a relatively inaccessible domain of particular geometrical complexity. While the complete zone is actively ventilated in order to reduce the build-up of any flammable vapour and to control ambient temperature levels, such flows will also exert a strong influence over the development of any fires that might arise from the loss of fuel or lubricant containment. The introduction of new materials, or re-sizing and relocation of accessories, together with enhanced use of electronic control systems and the accompanying tighter limits on the ventilation environment will tend to undermine any existing experience base, and emphasise the need to develop robust and flexible simulation tools. The combination of poorly controlled ventilation flows and complex physical obstacles in close proximity to each other makes the prediction of fire growth and of the accompanying rates of component heat transfer especially difficult. Standard fire safety assessments that are based upon simplifying procedures ­ such as those of whole zone heating to a prescribed temperature level or burner tests of the survivability of individual components ­ can be used to set conservative thresholds. Such coarse design constraints are increasingly incompatible with the introduction of finer thermal tolerances on particular electronic components or the introduction of novel materials. Increased reliance on the bulk zone ventilation to provide a suitable operating environment for a range of accessories also demands greater flexibility in their positioning. Such considerations can also contribute to improvements in their ease of assembly and subsequent maintenance. Computational Fluid Dynamics (CFD) is uniquely well-suited to the analysis of the interactions required by a more responsive approach to fire safe design. Although the methodology is widely employed in the aerospace field to assist in the development and implementation of specific technologies, the simulation of accidental occurrences poses a number of different challenges, especially in relation to validation. Design point component performance provides a clearly defined focus for the application of CFD and permits a measure of model calibration when applied to a complex component like the combustion chamber. Such considerations are generally absent in the analysis of fire where, for example, neither the level of fuel flow rate nor the point of delivery can be readily prescribed. Of course reliable CFD simulations for such complex scenarios cannot be achieved without supporting experimental validation. Simplified representations of the nacelle geometry have been incorporated in two, approximately half scale, experimental rigs. A Perspex rig for water analogy flow visualisation and a water-cooled, mild steel design, for fire event simulation.

Figure 2. Ventilation flow development within the casing inferred from water flow visualisation (View from bottom dead centre)

The basic features of the rig are illustrated in Figure 1 (opposite). An annular fire zone is created between two concentric cylinders. In order to permit repeated fire experiments without damage and distortion of the zone envelope, both cylinders are water-cooled. Within the annular space created between the cylinders a range of obstacles of varying size and shape has then been fitted, representing different engine accessories, for example oil coolers, gearbox and control units. The annulus is often particularly congested towards bottom-dead-centre (bdc) where the gearbox, starter motor and air starter duct are typically located. Ventilation air is introduced into the casing through a bifurcated inlet mounted at the top of the nacelle. The outer drum is ribbed ­ as would be the situation in practice with a stiffened thin-walled structure. Although a number of different types of fire source have been examined in this rig, including kerosine sprays ignited both locally and remote from the fuel source, we shall largely focus in this paper on pool fire scenarios. In these it has been assumed that liquid kerosine has pooled at the base of the zone, following a fuel supply leak and burns as a turbulent diffusion flame. No attempt has been made in these studies to simulate any particular ignition process. Pool fires arise in many practical scenarios and have therefore attracted considerable research attention within the fire engineering community. The confined casing fire

EXHAUST

GEARBOX

configuration is quite distinctive, however, displaying features that are not encountered in other applications where the effects of either buoyancy or forced convection predominate. The ventilation flow and that of the fire plume are, in principle, opposed but, since the internal accessory layouts are not symmetrical, plume development occurs in only half of the rig and may therefore be assisted by the ventilation flow spilling over from the opposite half.

Fire Plume/Ventilation Flow Interaction

The introduction of high momentum air jets into the large annular space, even in the absence of fire, creates a particularly complex, maldistributed flow field. Figure 2 illustrates typical isothermal flow simulations of a minimally congested zone, based on observations from the water analogy experiment performed with a Perspex model. The two jets quite clearly do not separate the zone into weakly interacting left and

Figure 1. Experimental rig ­ typical configuration and accessory layout

14

Aerogram Bringing News from the Horizon

June 2006 Volume 11 Issue 6

15

Gas turbine engine nacelle ventilation and fire events... (continued)

Computational rotor aerodynamics at Cranfield University

Dr Scott Shaw, Department of Aerospace Sciences, School of Engineering

right halves. A significant part of the clockwise directed inlet jet is deflected to the front of the rig and completes a full circuit of the zone. Such behaviour has important implications for the plume trajectory in the event of fire and the nature of fire attack on particular zone accessories. Key elements in the numerical simulation of these scenarios are evidently the plume trajectory, and its interaction with the forced ventilation, the mixed plume temperature levels and the associated heat fluxes to critical accessories.

clockwise direction and to wrap around the first internal obstacle (representing an oil cooler). At the flow rates and heat release rate simulated the plume does not reach top dead centre of the rig. Figure 5 presents a more detailed visualisation of the flow streamlines in the vicinity of the plume and the fire source. Here small scale geometric details are seen to be important, specifically the proportion of unburnt fuel that `short circuits' to the exhaust, by flowing below the obstruction of the gear box. Finally, Figure 6 presents a surface plot of the local convective and radiative heat fluxes to the interior surface of the nacelle. Unsurprisingly the peak incident fluxes are in the region of the fire plume around the gearbox and oil cooler, with a maximum value of approximately 30 kW/m2. While some uncertainty remains over the local gas temperature, the absolute values must be interpreted cautiously, however the indicated distributions are not expected to change significantly.

Numerical computation of the flow around a full helicopter configuration (main rotor, fuselage and tail rotor) presents a number of significant challenges for rotor aerodynamicists. In both hover and high-speed forward flight the disparate range of length and time scales of the flow provides a rich aerodynamic environment that is dominated by discrete vortical flow features, boundary layer separation, shock-motion and transition. That these phenomena occur at Mach numbers, incidence and Reynolds numbers that change significantly more than ten times a second only further confounds theoretical approaches. Despite these difficulties there has been rapid progress in the past decade towards the practical numerical simulation of a complete helicopter configuration. At Cranfield our initial focus was on the steady and unsteady aerodynamics of aerofoils over the range of Mach numbers, lift coefficients and Reynolds numbers that characterize the rotor operating envelope. Effort was also expended developing robust numerical treatments of the rotor wake in hover which have resulted in a sophisticated approach based upon a priori grid adaptation and high-order numerical schemes suited to vortex wake capturing. More recently our effort has been directed towards the calculation of forward flight of isolated main rotors and fuselages and their aerodynamic interaction. In this article we use recent results to illustrate the important physical phenomena experienced by rotorcraft in hovering and forward flight and review progress towards the provision of a practical tool for the aerodynamic prediction of a complete helicopter in hovering and forward flight.

Computational Simulation

Numerical simulations have been carried out using both the CFD code, SOFIE (Simulation of Fires in Enclosures), specifically developed at Cranfield for compartment fire predictions, and also the commercial code Fluent. Given the earlier observations regarding flame development and impingement, one particular focus has been related to the relative length and local temperature of the fire plume. Once confidence has been demonstrated in predicting local temperatures then a more quantitative evaluation of radiative and convective heat fluxes to the individual components within the nacelle may be carried out. Figure 3 presents the predicted flow streamlines within the rig in the event of a pool fire and illustrates the considerable interaction between the complex fluid dynamics of the designed ventilation system and the fire plume. Whilst the level of congestion is here quite modest and the rig represents only a generic geometrical configuration, the complexity of the flow is clearly visible. In particular, elements of the ventilation flow clearly do not pass immediately to the exhaust but rather circulate one or more times before finally escaping. Even without considering the fine detail of the fire simulation, the potential of CFD for optimising the ventilation system is evident. Figure 4 provides a visualisation of the fire plume through an iso-surface of mixture fraction (which may be considered to be proportional to the local air/fuel ratio). The plume is seen to be deflected in a

Figures 3 - 6, top to bottom above Figure 3: Visualisation of predicted flowfield in event of a pool fire (streamlines coloured by velocity magnitude). Figure 4: Visualisation of predicted fire plume (plume coloured by gas temperature). Figure 5: Near field view of flow streamlines around fire plume. Figure 6: Predicted incident heat flux to casing walls and accessory units.

Conclusions

The interaction between the ventilation flow in a representative engine nacelle compartment and the pool fire that may result from the loss of containment of fuel or lubricant, has been simulated both experimentally and computationally. The trajectory of the fire plume is revealed to be highly dependent upon the layout of accessories in the zone and this, in turn, affects the local heat transfer environment. Temperature measurements on obstacles representing different accessories within the zone reveal substantial temperature differences depending on the fire plume trajectory. Numerical simulations demonstrate the potential for CFD to simulate fire events within gas turbine engine nacelles. Key flowfield features are readily reproduced and provide opportunities for the tailored design of the overall ventilation flow and the siting of the various accessories.

Figure 1: Unsteady aerodynamic environment experienced by a rotor in fast forward flight. (a) Local Incidence (degrees). (b) Local mach number.

Unsteady aerodynamics

Unlike fixed wings, the aerodynamic environment of rotors is intrinsically unsteady. Unsteadiness arises from a number of sources. The varying component of the forward flight velocity experienced normal to the blade, cyclic control inputs and blade flapping all have large amplitude variations that occur at the frequency of the blade rotation. Typical behaviour for a helicopter rotor in fast forward flight is shown in Figure 1 which plots the local Mach number and incidence experienced within the rotor disc. There are additional high frequency loads arising from the interactions of the rotor with the wake system generated by preceding blades and with the airframe. If the helicopter rotor blade is conceived as a high aspect ratio wing (typical aspect ratios are in excess of 15) then it is possible to justify the use of a twodimensional approximation and the blade loads can then be calculated by the use of a strip theory in which the local

aerodynamic loading is obtained by considering the aerodynamics of individual blade elements at appropriate Mach number and incidence. A typical locus of Mach number and incidence for an individual blade section (here at the tip of the Westland Wessex flying at its maximum flight speed) is shown in Figure 2. While both the variation of Mach number and the incidence may be responsible for significant unsteady aerodynamic effects in fast forward flight the complexity of the problem is such that it is often convenient to consider the aerodynamics of the individual motions in isolation.

Acknowledgements

The research reported has been undertaken with financial support from Rolls-Royce plc and the UK Department for Trade and Industry. Their support and interest is gratefully acknowledged. The water analogy study, underlying Figure 2, was performed by Mr D Binks.

Incidence variation

The aerodynamics of aerofoils undergoing pitching oscillations have been widely investigated because of the importance of the physical phenomena for a number of aerospace engineering problems. For attached flow conditions the force and moment hysterisis arising from the unsteady

Figure 2: Flow conditions experienced by aerofoil in forward flight

16

Aerogram Bringing News from the Horizon

June 2006 Volume 11 Issue 6

17

Computational rotor aerodynamics...(continued)

Local Lift

Figure 3: Computed and measured air-loads for a pitching aerofoil. (a) Lift coefficient (b) Drag coefficient

Span

Time

motion can be calculated with a high degree of confidence, Figure 3. The observed hysterisis occurs as a consequence of two effects; the generation of acoustic waves by the piston-like motion of the aerofoil through the air and the convection of vorticity that is shed into the wake. At higher incidence dynamic stall occurs. Dynamic stall is a complex physical phenomena involving the formation, shedding and motion of large scale vortical structures. A typical sequence of events is illustrated in Figure 4. Initially the flow is fully attached, but as the aerofoil pitches up beyond the static stall angle of attack a stall vortex forms at the leading edge. This vortex allows the aerofoil to continue generating lift beyond the value that can be obtained for a stationary aerofoil. At the top of the up stroke the stall vortex separates from the leading edge and is then convected downstream over the aerofoil surface. The induced effects of the vortex as it traverses the aerofoil are significant, leading to rapid variation of the magnitude and sign of the pitching moment coefficient. The flow then remains separated until the incidence is well below the static stall incidence. The amount and shape of the aerodynamic hysterisis is highly non-linear depending on the shape of the aerofoil, the Mach number, the Reynolds number and the amplitude and frequency of the motion. The sensitivity of the physical phenomena poses a number of challenges to current CFD methods that require continued research in the areas of turbulence and transition modelling.

effect of such oscillations is poorly understood due to the difficulty of studying such flows experimentally. From a computational perspective the simulation of the required in-plane motion is no more difficult than that of pitching motion, and so in the mid 90s Cranfield pioneered the study of in-plane oscillations using Computational Fluid Dynamics (CFD). Below the Mach number at which shock waves appear the effects of in-plane motion on blade loading can largely be described using steady aerofoil performance data with an appropriate phase lag, but once the section begins to generate shock waves the non-linear nature of the subsequent flow development leads to a number of unexpected consequences. Figure 5 presents computational Schlieren pictures over the advancing side of a nonlifting rotor. Comparing images at constant nominal Mach number (azimuth angles that are symmetric about 90° azimuth) the effects of flow hysteresis are clearly evident. As the aerofoil accelerates, the formation of the shock wave is delayed to Mach numbers beyond that at which they would normally be expected, while conversely as the aerofoil decelerates at azimuth angles beyond 90°, the shock remains much longer than would normally be expected. The resulting shock motion cannot be predicted from the usual quasi-steady considerations, indeed as a result of flow acceleration and deceleration the shock may occur at positions that cannot be attained under steady conditions. Understanding this behaviour may provide the key to increasing the forward flight speed of helicopters as control loads generated by the unsteady shock and the related movement of centre of pressure can exceed the safe operating limits of the control system thus limiting the maximum attainable forward flight speed.

Figure 6: Indicial response of a swept wing to an impulsive change in incidence.

Cn

Figure 7: Comparison of an indicial model and measured data for pitching motion.

Figure 8: Contours of local lift during low speed descending flight.

Reduced order modelling

While simulations such as those presented in the preceding sections provide a detailed knowledge of the flow at any instant in time and any position in space they may not provide the practicing engineer with physical insight or design awareness, indeed the sheer scale of the data provided from a single calculation may be overwhelming. There is therefore a clear need to find ways of distilling the data from full simulations so that the information can be employed within lower-order models that can then serve as the basis for further analysis. By doing this we hope to retain the fidelity of the underlying simulations but provide a tool that produces the results required by the design engineer in a fraction of the time. In modelling the aerodynamics of helicopter rotors techniques based upon indicial theory have been widely used in conjunction with experimental data to provide reduced order models of the unsteady forces and moments generated by the blade during its motion. At the heart of this method are the indicial responses of the system, in this case the aerodynamic response of the forces and moments to impulsive motions of the aerofoil. These responses are a mathematical conceit, impulsive motion cannot be reproduced in a wind tunnel experiment, and must therefore be determined analytically or indirectly from experimental data. Computational fluid dynamics provides a means of determining indicial response functions directly and at Cranfield we have developed tools and methods for the extraction of indicial response functions from CFD computations for simple unsteady motions including pitch, heave, in-plane

and gust responses. The normal force response of a three-dimensional swept wing to an impulsive change in incidence is shown in Figure 6. The form of the lift response reveals the physical mechanisms that are responsible for the force and moment hysterisis observed during more complex motions. The response is made up from two basic contributions; a large initial pulse that rapidly reduces in amplitude and a more gradual build up of lift towards a final steady value. The initial impulse loading is related to the generation of acoustic disturbances caused by the displacement of fluid by the aerofoil. While the later stages of the response are associated with a vortex that is shed into the aerofoil wake. This vortex is similar to the classical starting vortex generated by lifting surfaces and acts to reduce the incidence experienced by the aerofoil. As the vortex is carried downstream of the aerofoil its influence diminishes, allowing the effective incidence to recover to its final steady value. Once the response to a step input has been determined, the behaviour of the forces and moments to any arbitrary motion schedule can be computed using linear superposition. Figure 7 compares the lift response obtained from a CFD based indicial model with experimental data for an aerofoil performing simple pitching oscillations. The fidelity of the approach is comparable with that of conventional CFD methods but the utility of the approach becomes obvious when one considers that the computational time required by the indicial model is measured in fractions of a second. Indeed, the simplicity of the model is such that it could be used to generate forces

and moments for arbitrary motions in real time. The indicial model can be readily extended to three-dimensions for high-aspect ratio rotor blades to provide a comprehensive model of the aerodynamics of forward flight. In Figure 8 the distribution of lift generated by the main rotor of the Westland Lynx during low speed approach has been computed using GKN Westland's Helicopters ACROT comprehensive model in conjunction with an unsteady aerodynamic model derived from CFD that includes the influence of Mach number and incidence variation.

Figure 4: Visualisation of dynamic stall events on NACA 0012 aerofoil.

Transition modelling

For most of the past 80 years, the harsh aerodynamic environment in which rotors operate has led rotor aerodynamicists to believe that the boundary layer on rotating wings could be considered fully turbulent. This observation appeared to be supported by measured power coefficients which could be predicted to a high level of accuracy using data for fully turbulent aerofoils. More recent laboratory experiments on rotating wings and data collected during flight test campaigns suggest that this assumption is erroneous and shown that regions of laminar flow may extend over significant regions of the blade. These observations are of particular interest as envelope limiting aerodynamic phenomena during forward flight, such as dynamic stall, are known to be highly sensitive to transition physics. Clearly there is a need to improve the modelling of the boundary layer on helicopter rotors by including transition physics if we are to provide better estimates of power consumption and stall

Mach Number Variation

In forward flight the rate of advance of the helicopter provides a significant cyclical perturbation to the velocity experienced by individual elements of the blade. In contrast to incidence variation, the unsteady

Figure 5: Instantaneous Schlieren Images of flow around aerofoil during in-plane motion.

18

Aerogram Bringing News from the Horizon

June 2006 Volume 11 Issue 6

19

Computational rotor aerodynamics...(continued)

Transition Location (x/c)

Figure 9: Comparison of computed and measured transition behaviour.

Cp

Figure 10: Comparison of computed pressure and skin friction with experiment (M=0.15, Re=300,000).

Figure 14: Prescribed surface grid for a twobladed hovering rotor. Figure 15: Cross section through the adapted grid and solution.

Figure 16: Computed and measured spanwise lift distribution.

limits. In addition detailed computational studies may also help to improve our knowledge and understanding which will ultimately result in improved rotor design. At Cranfield we have sought to address this problem through the incorporation of empirical models of the transition process within our existing viscous flow model. To this end we have evaluated a range of empirical criteria and selected those that provide both accuracy and reliability across the range of flow conditions experienced by rotorcraft. The resulting model performs well for steady flows, see for example Figures 9 and 10 in which computed (solid lines) and measured (symbols) transition data are compared for two-dimensional low Reynolds number flow over a wide range of incidence including stall. For unsteady flows the model leads to non-trivial improvements in the predicted air-load hysterisis for attached flows and also improves the qualitative behaviour of the simulations under dynamic stall conditions. However a lack of reliable experimental transition data for such flows inhibits a rigorous evaluation of the model. More recently the ability of the model to compute transition location for an isolated rotor in hover has been assessed, see Figure 11. Preliminary results are encouraging but further refinement is required before the model can be used in an engineering design environment.

xt

rotor. The ability to accurately calculate the influence of the wake generated by an isolated rotor and associated interactions is therefore of great practical importance in understanding and improving blade aerodynamic design. In hovering flight the rotor wake comprises two key physical elements: a sheet of trailed vorticity associated with span wise variations of blade loading and a pair of concentrated vortices trailed from the blade tip and root, see Figure 12. The vortex sheet is a relatively weak feature of the flow that descends in a tightening helical pattern below the rotor. The root and tip vortices follow contracting helical trajectories below the rotor disc. The tip vortex is generally much stronger than the root vortex due to the significant difference in dynamic head experienced at the tip and hub. The initial development of the wake sheet is thought to be important in establishing the rotor inflow, but at later wake ages, in the `far wake', the tip vortices dominate the flow development. The physical characteristics of the rotor wake, in particular the importance of concentrated regions of vorticity are particularly challenging for current computational methods. The predicted wake of a two-bladed rotor in hovering flight is shown in Figure 13. The main features of the flow are resolved in the computation. The trajectory of the vortices generated at the tip and root of the blades are clearly visible as is the wake sheet trailed behind the blade. The fidelity of the current method represents an improvement over earlier numerical attempts to compute the rotor wake and is principally related to the use of an a priori wake adaptation technique in which an empirical prescribed wake model is used to estimate the location of important

flow structures such as the tip vortex and wake sheet, see Figure 14. This information is then used to create additional computational surfaces in the grid that have the effect of refining and aligning the computational grid around the flow structures, Figure 15. This a priori wake adaptation technique permits high-fidelity computations to be performed using a fraction of the grid points required by conventional meshing techniques. The predicted wake development shown in Figure 15 required approximately 2 million nodes per blade which compares favourably with the 32 million nodes reported by other investigators to achieve the same level of accuracy using more conventional structured grid approaches. Clearly the ability to use grids containing fewer cells has implications for the cost of individual calculations, both in terms of the overall CPU time consumed and in the time to solution. The ability to properly resolve the structure of the near-wake allows the blade loading and ultimately rotor performance to be evaluated with a high degree of confidence. Figure 16 presents comparisons of computed and measured spanwise blade loadings that illustrate the accuracy achieved in the current simulations.

Figure 11: Comparison of computed and measured transition locations.

Figure 17: Computed pressure contours at various blade azimuths.

Figure 12: Illustration of the wake of a single bladed rotor in hover

maintains time accuracy. This can be achieved by rotating the entire domain at the required angular velocity. This method has the advantage that it requires no special treatment within the solver. Results for a two-bladed non-lifting rotor in fast forward flight are shown in Figure 17. The results show surface pressure distributions at a number of azimuth stations. The azimuthal development of shock waves at the blade tip, which has a maximum incident Mach number of 0.8, can be clearly observed as the blade advances in the direction of the oncoming flow. The thrust generated by the same rotor system at a lower forward flight velocity and 8 degrees of collective are compared with simple blade element theory in Figure 18. In the current calculations the blade is considered to be rigid resulting in a thrust history that has a frequency twice that of the rotor. In light of the previous discussion this assumption is clearly unreasonable. Coupling the CFD solver with a model of the blade structural dynamics presents a number of challenges and is currently being addressed in conjunction with the University of Sheffield and Aston University.

Rotor Fuselage

Aerodynamic interactions between the main rotor and fuselage are responsible for a number of significant problems in the design of rotorcraft. The effect of the rotor wake impinging on the fuselage is responsible for significant vibratory loads and noise, while the resulting unsteady downwash can compromise the stability of the vehicle. The airframe also impacts the performance of the main rotor and tail rotor systems as the presence of the fuselage changes both the wake trajectory and the inflow distribution. The fluid dynamic mechanisms contributing to these aerodynamic interactions are very

Forward Flight

In forward flight, as we have already seen in two-dimensions, the component of forward flight velocity experienced normal to the blade leading edge introduces significant flow unsteadiness. In contrast to the hovering rotor, where it is possible to transform the problem to a steady flow by choosing a frame of reference attached to the rotating blade, we must now compute the flow in a manner that

Thrust coefficient Figure 18: Comparison of computed data and simple theory for forward flight.

Hovering flight

Unlike fixed wing aircraft the helicopter rotor operates in close proximity to its own wake resulting in significant interaction. The induced effects of such interactions have a substantial influence on blade loading, aerodynamic performance, vibration and aero-acoustics of the main

Figure 13: Computed wake of a two-bladed rotor in hovering flight

20

Aerogram Bringing News from the Horizon

June 2006 Volume 11 Issue 6

21

Computational rotor aerodynamics...(continued)

complex and today many aspects of the problem still remain poorly understood. As a step towards understanding this problem more fully Cranfield University has begun investigating the NASA ROBIN configuration. This main rotor and fuselage configuration is well documented and has an extensive dataset for the fuselage and rotor alone and the fuselage and rotor in combination. Computed contours of static pressure for a modest forward flight Mach number are shown in Figure 19 for the body alone, while Figure 20 illustrates the comparison that can be achieved with the measured data. Work computing the fuselage rotor combination using sliding and deforming meshes is at an early stage.

reducing the acoustic signature of the resulting blade vortex interaction. The CFD tools described in the earlier part of this article have been used to study the aerodynamics of this and alternate configurations that achieve the same aim. The study of such configurations provides an opportunity to understand and explore the use of CFD as both a rotor analysis and design tool. Indicial models of the pitch and gust response of a two-dimensional section generated using Cranfield's twodimensional code have been used to optimise the spacing of the vortex trajectories so as to maximize the constructive interference between the vortices. This is illustrated in Figure 21 (opposite) which shows the rate of change of lift with time of the original configuration and the optimised configuration. The time rate of change of lift is responsible for a significant part of the noise signal and the present results suggest a substantial reduction. The resulting design was subsequently tested at low speed as a fixed wing, the wind tunnel model and a mapping of the vortices is illustrated in Figure 22 (opposite). The experimental data clearly show that the vortex dynamics are more complex than those assumed in the simple model, the twin vortices tend to orbit one another as they travel away from the blade tip. Although the gross effects of the twin vortex system can be realised the subtleties of the wake dynamics would appear to play an important role in the

overall system performance as the relative orientation of the vortices changes with time. This has also been demonstrated computationally for the case of a rotor in hover, Figure 23. For this case the twin vortex system is established using blade dihedral, the usual tip vortex is seen together with a second vortex from the junction. Using Cranfield's a priori grid

adaptation technique we have been able to capture the complex dynamics of the tip and junction vortices as they orbit one another. This information is currently being used in conjunction with the indicial model to understand the potential impact on acoustic loading.

Conclusions

In this article we have used the computational capability developed at Cranfield for rotorcraft to illustrate the aerodynamic complexities and difficulties faced by helicopter aerodynamicists. While the main concentration at Cranfield has been on the development of tools, processes and techniques that provide high-fidelity analyses we have also sought to consider how these tools, and the insight that they provide, can be used effectively within a rotorcraft design environment. The ability to compute steady and unsteady flows around aerofoils over the wideranging conditions experienced by rotorcraft has been demonstrated along with our capability for isolated rotors in hovering and forward flight and isolated rotor fuselages. In our most recent work we have begun to explore the full complexity of the rotor-fuselage interaction problem using our unsteady solver in conjunction with sliding boundary planes and a priori grid adaptation, Figure 24. Once we have understood and demonstrated the reliability of this approach we anticipate using the resulting tool to explore how the performance of rotor systems can be improved through refinements in aerodynamic design, some initial steps in this direction using multi-vortex systems generated through simple planform variations were presented.

Figure 19: Computed contours of static pressure on ROBIN fuselage.

Aerodynamics of Novel Tip Shapes

A major barrier to the public acceptance of helicopters within the urban environment is rotor noise. There are a number of physical mechanisms responsible for rotor noise, but perhaps the most familiar and intrusive is that due to `blade slap' or blade vortex interaction (BVI). The literature abounds with techniques for reducing BVI including the vane tip device proposed by the aerodynamics group at Westland Helicopters. This elegant solution aims to produce two vortices rather than the single tip vortex of a conventional design. This has several advantages; firstly the individual vortices are weaker and secondly by careful design the vortex trajectory can be manipulated to provide interference between the vortices,

Figure 21: Optimisation of vortex spacing.

Figure 22: Wind tunnel model and vortex map for optimised vortex spacing.

Figure 20: Comparison of computed and measured surface pressure.

Cp

Cp

Cp

Figure 23: Wind tunnel model and vortex map for optimised vortex spacing. ACKNOWLEDGEMENTS The contributions of former member of staff Professor Ning Qin (now Sheffield University), PhD students Dr Jason Hill, Carlos ManglanoVillamarin and Katherine Knight and MSc students Dorothee De Villelle, A Jedar, Marcus Weiber, Jorge Estevez Garcia and Alex Berlioz to the work reported in this article are gratefully acknowledged. Aspects of the work were made possible through the financial support of the Engineering and Physical Science Research Council (EPSRC) and Westland Helicopters Ltd. The three-dimensional rotor and fuselage calculations were performed on the Cambridge-Cranfield High-Performance Computing Facility located at Cranfield University.

x/R=0.0517

x/R=0.145

x/R=0.2563

Cp

Cp

Cp

x/R=0.3498

x/R=0.6003

x/R=1.007

Figure 24: Unstructured mesh and sliding plane topology for Rotor-Fuselage calculation.

22

Aerogram Bringing News from the Horizon

June 2005 Volume 11 Issue 5

23

Integrating CFD and experiments in aerodynamics

Dr Kevin Knowles and Dr Alistair Saddington, Department of Aerospace, Power and Sensors, Defence College of Management and Technology

This article discusses how the combined use of computational fluid dynamics (CFD) and experimentation can give additional insight into aerodynamics problems. Examples from three particular areas of research at Shrivenham are given: transonic cavity flows, high-speed turbulent jet flows and open-wheeled race car aerodynamics. In each case, knowledge gathered from one analysis technique has been used to assist in the application of the second technique thereby enabling greater understanding of the flow physics being studied than would have been possible through the isolated use of one or other methodology.

Figure 2: PIV results for three planes across the width of a L/D = 5 cavity at M = 0.85 (Z/W = 0.5 is the centreline).

Figure 3: Particle pathlines from CFD for L/D=5 cavity at M=0.85. Half-domain calculated on a Linux cluster using 952000 cells (mirrored for full cavity view) with the Realizable k-e turbulence model.

Transonic cavity flows

Transonic flows over rectangular cavities are of particular relevance to aircraft weapons bays, Figure 1 (opposite). For cavity length-to-depth ratios less than about 10 the flow is characterised by intense acoustic tones and flow unsteadiness. We present here numerical modelling and particle image velocimetry (PIV) measurements of a transonic cavity (M=0.85) with a length-to-depth ratio (L/D) of 5. Here, one use of the numerical model is to aid optimisation of the PIV set-up.

Numerical model

Figure 1: F-22 showing weapons bay (courtesy of Lockheed Martin).

Introduction

CFD practitioners and experimentalists have a common goal of understanding aerodynamics. It is therefore surprising that often the disciplines have only a very limited interaction, which usually involves the validation of CFD. This practice supports integration at the most basic level but often there is no interaction whatsoever between the experimentalist and the CFD practitioners. This situation is unsatisfactory from many points of view including: · the need to have an appreciation of the flow before deciding what should be measured · the desirability to check experimental measurements as they are taken · the difficulty in making certain important measurements · the need to assess the influence of the experimental techniques on the measurements · the ability of CFD to provide detailed flow information and sensitivity at a reasonable cost for some cases · the large cost of CFD calculations for other cases · the lack of credibility for the CFD results for some flow categories.

It could be argued that the process of aerodynamic investigation would be significantly enhanced if the integration of CFD and experiments were much stronger. In particular, the design and reliability of experiments could be significantly enhanced by CFD, the scope of experimental measurements extended through CFD and the credibility of the simulation results enhanced by the availability of suitable measurements from experiments. This sort of closer integration is, however, rare. To encourage more careful consideration of these issues and to stimulate new ways of approaching aerodynamic studies, Cranfield University at Shrivenham hosted the Second International Symposium on Integrating CFD and Experiments in Aerodynamics on 5th and 6th of September 2005. For some time now, aerodynamics research carried out by Cranfield University at Shrivenham has made use of both experimental and numerical methods. The combination of these two disciplines has enabled greater understanding of the flow physics being studied than would have been possible through either a purely experimental or purely numerical approach. Examples of some current research projects which use this integrated approach will now be discussed.

The 2-D CFD model was completed using commercial software GAMBIT + Fluent 6. The computational grid was created from 86000 uniform quadrilateral cells with a distribution of 320 x 65 cells within the cavity. The domain was designed to be geometrically similar to the wind tunnel test section. Thus, the upper and lower surfaces of the tunnel were modelled with wall boundary conditions while the inlet and outlet conditions were modelled as the pressure far-field or free air condition. The flow problem was solved using the unsteady coupled solver and realisable k- turbulence model. The k- family of turbulence models was selected because a number of previous studies have shown it to perform well for flows with high shear and regions of recirculation.

Experimental model

5mm thick glass and is mounted in a modified wall in our transonic wind tunnel. Due to the design of the tunnel test section it was not possible to image the entire cavity and into the freestream so the top 3mm inside the cavity could not be mapped due to the presence of a flat metal plate. An in-house-designed seeder injected a 5% glycerol and water solution into the contraction section of the wind tunnel for the PIV measurements. PIV images were recorded in sets of 70 images for each run. Data processing was carried out using the TSI software UltraPIV. The performance of the UltraPIV algorithm, however, was found to be poor in regions of low seeding density, so an in-house algorithm was developed based on correlation-averaging, Figure 2.

Experimental optimisation

experiment for use in transonic and supersonic flows. In any given experiment, other parameters critical to the PIV system's performance include the seeding density and corresponding particle image density, Ni, and the interrogation region size relative to particle image size, D/d. All these parameters must be considered when optimising the correlation performance. With a priori knowledge of Dr, ø and Vmax, D/d and Ni, it is possible to use the method to determine a laser pulse separation, t, for a PIV experiment which will ensure that, on average, from a series of experiments at least 50% of valid vectors will be obtained from a `worst case' interrogation region in the flow. The experiments are then defined as being optimised. The CFD predictions provide the a priori knowledge of Dr, ø, Vmax and Vmin, where the spatial resolution requirement will determine the parameters D/d and Ni from the PIV experiment. The optimisation method will then yield a recommended magnification, M and pulse separation, t.

Results

PIV data was taken from inside an allglass cavity of 160mm length, 80mm width and 32mm depth. The cavity is made from

Lawson et al. [1] have described an optimisation method for a double-pulsed PIV experiment which can be applied to autocorrelation or cross-correlation analysis. They have shown that in order to retrieve a valid velocity vector from an interrogation region there exists a strong interdependence between the dynamic range, Dr of the flow, defined by Dr=Vmax/Vmin (where Vmax and Vmin are the maximum and minimum velocities measured in the flow plane) and the velocity gradient strength, ø, defined by ø=(V2 - V1)/Vmax. Here, V2 - V1 is the velocity change across the interrogation region. A high dynamic range requirement necessarily restricts the strength of the velocity gradient in a chosen region and vice versa. The latter condition is crucial to the design of a PIV

By comparing the output of the PIV with the CFD predictions it was found that optimisation of a given PIV system requires not just a-priori knowledge of the flow to specify variables such as M and t, but also careful control of variables such as seeding and judicious choice of the data processing algorithms for a given flow. Currently, CFD is being used to help understand the complex three-dimensional, unsteady flows within cavities, Figure 3.

24

Aerogram Bringing News from the Horizon

June 2006 Volume 11 Issue 6

25

Integrating CFD and experiments in aerodynamics...(continued)

Numerical model

The CFD model was developed using the Fluent commercial code (Version 5.5). The computational domain consisted of a threedimensional hexahedral mesh with some 250000 to 450000 cells depending on the test conditions. Only one quarter of the geometry was modelled since each nozzle has two axes of symmetry. The inlet plane was approximately 1D up-stream of the nozzle exit, the outlet plane was at about 50D downstream and the radial boundary diverged from 2D at the upstream end to more than 10D downstream. The boundary condition for the nozzle inlet was set as a pressure inlet with a prescribed total pressure, static pressure, total temperature and turbulence intensity. The turbulence intensity, Ti, at nozzle inlet in the CFD model was adjusted to give the same nozzle exit turbulence intensity as the experiments (approximately 4%). The experimental Ti was derived from the rms velocity data measured by the LDV technique. The turbulence length scale was set as 7.5% of nozzle radius. The far field boundary was set as a pressure outlet with a prescribed static pressure and static temperature. Turbulent calculations were performed using the RNG k- turbulence model, which has been shown to be suitable for modelling under-expanded jets.

Mixing enhancement

in mixing rates lower than an axisymmetric jet with final mass flow rates similar in each case. Although the mixing enhancement could be determined experimentally, the entrainment mechanism responsible could only be determined from the CFD.

were made in vertical planes oriented perpendicular to the freestream flow. When processed, these data subsequently provided time-averaged velocity results.

Numerical model

Open-wheeled racecars

a) Ground simulation and wheel rotation are known to be essential for accurate automotive testing, particularly for openwheeled racecars. With this type of vehicle, ground effects and large unfaired wheels dominate their aerodynamic characteristics. Every care must be taken to ensure that the wheels are modelled correctly both in experimental testing and computational simulation. Previous evaluations of the capability of CFD to model wheel flows used the surface pressure and force data published by Fackrell [3] as their main validation criteria. CFD models enable researchers to investigate physical processes that may be impossible to reproduce experimentally. In our work [4] the investigation concentrated on the effect of using external wheel support struts during racecar wind tunnel testing. The struts are used to mechanically decouple the car body from the ground while still allowing the wheels to rotate. Ideally the support struts should be aerodynamically neutral, however, this is not always the case. An experimentally validated CFD model of an isolated racecar wheel and strut was used to quantify the aerodynamic interference effects between the two. The virtual environment of the CFD model enabled the support strut to be easily removed, something that could not have been carried out experimentally.

Experiments

Figure 4: CFD prediction of a rapid-mixing nozzle, showing distortion of the jet cross-section due to differential expansion between castellations.

High-speed jet research

The rate at which a supersonic jet mixes with the surrounding ambient fluid is important for many aerospace applications, including controlling the infra-red signature of a propulsion plume. For supersonic jets the generation of streamwise vortices appears to be beneficial in improving mixing and various schemes have been investigated using vortex generators, tabs and other intrusive devices. Our work presents a numerical and experimental study of mixing in underexpanded, supersonic turbulent jets issuing from axisymmetric and castellated nozzles into quiescent conditions [2]. Experimental measurements using laser Doppler velocimetry (LDV) and Pitot probe measurements along the jet centreline at a nozzle pressure ratio (NPR=nozzle total to ambient static pressure ratio) of 4 indicated that the castellated nozzles entrained more mass flow into the jet than a simple convergent nozzle. The CFD models verified that this was the case but crucially also provided a physical explanation of the entrainment mechanism that would have been very difficult to deduce from the available experimental data.

Experiments

b)

Figure 5: An isolated racecar wheel and the external support sting used in wind-tunnel testing a) Model wheel and support sting in 1.5m wind tunnel. b) CFD grid of sting and wheel, including detailed hub modelling.

The wheel and sting assembly was placed in a rectangular domain with the inlet 5 wheel diameters upstream, outlet 16 wheel diameters downstream, a width of 10 and a height of 5 wheel diameters. The interior of the sting was also meshed to allow it to be removed from the solution domain by allowing fluid to flow through it, thus eliminating the need to generate an entirely new mesh for that section of the study. The only significant deviation from the experimental geometry was made at the tyre contact patch. Difficulty was encountered in maintaining high cell quality when modelling the near line-contact between the rolling road and nondeformable tyre. Therefore, the wheel was slightly truncated by raising the ground plane by 0.8mm. This increased the size of the contact patch and greatly improved the cell skewness in this area. The final mesh was of the order of 0.93million cells. The boundary conditions of the CFD simulation were chosen to be representative of those of the experiment. A uniform flow was specified at the inlet and standard atmospheric pressure specified at the outlet. The rolling road and wheel components were modelled as translating and rotating walls respectively. When simulating the wheel and sting, the sting surface was specified as a wall with the no-slip condition applied. When testing the wheel without the sting, the latter's surface was represented by an interior condition, which did not impede flow. The mesh inside the sting was solved as a fluid, effectively removing the sting from the domain. Symmetry planes represented the remaining domain boundaries. Simulations were run with the k- turbulence model.

Validation

Both the experimental measurements and the CFD model indicated that the castellated nozzles produce jets with shorter shock cells than the axisymmetric jet, however, it was still to be determined if this translated into enhanced jet mixing. For the purposes of the study, jet mixing was determined by a numerical integration of the mass flow rate passing through planes normal to the jet axis at various streamwise positions. The CFD model showed that the increased jet mixing produced by the castellated nozzles appeared to be due to differential expansion of the jet fluid in the gap and tooth regions as it leaves the nozzle exit, Figure 4. This differential expansion created a distorted jet cross section which presented a larger surface area to the ambient air, thus enabling more rapid entrainment. The mixing enhancement was, however, confined to the nearfield flow (x/D<10). At greater streamwise distances viscous dissipation appeared to cause the entrainment mechanism to decay resulting

Experiments were conducted in a nozzle test cell at Shrivenham. Three-dimensional LDV measurements were taken of the flow from the castellated nozzles. Measurement traverses were made along the nozzle centreline, x. Probe access limited data collection to the first 10D (ten nozzle diameters) from the nozzle exit plane. The LDV measurements were estimated to be accurate to ±1% of velocity based on the sample time and frequency.

26

Aerogram Bringing News from the Horizon

A 40% scale (263mm diameter), nondeformable Champ Car front wheel assembly was chosen along with its associated support sting, Figure 5. Additional tests were also conducted on a 50% scale F1 wheel, together with three different support stings. LDV measurements were made using a two-component and a single-component, 1m focal length, Dantec FibreFlow probes mounted to a three-component traverse. Data acquisition was carried out by three BSA enhanced signal processors, with all equipment centrally controlled by Dantec Burstware software. LDV measurements

Figure 6: Contours of static pressure on wheel and sting derived from CFD.

The mean drag force calculated from the experimental data was non-dimensionalised by the frontal area of the wheel. The drag coefficient, CD predicted by the CFD simulation was 0.638, which was 6.2% lower than the measured value of 0.680. Care should be exercised when using force data as the sole accuracy measure but the good experimental velocity correlation supported the validity of this prediction. This was further reinforced by

June 2006 Volume 11 Issue 6

27

Integrating CFD and experiments in aerodynamics...(continued)

CFD Data ­ With Sting ­ 54% Plane

CFD Data ­ Without Sting ­ 54% Plane

the increase in lift. Flow through the wheel is from outside to inside, therefore the results suggest that the sting forces more flow to pass through the wheel than would otherwise occur, see Figure 8. In the case of the F1 wheel, flow is the other way and the sting was found to impede this. This study has shown how an experimentally validated CFD model of an isolated race-car wheel and strut can be used to quantify the aerodynamic interference effects between the two. The virtual environment of the CFD model enabled the support strut to be removed easily, something that could not have been done experimentally.

CFD in aerospace design

Professor Mark Savill & Dr Karl Jenkins, Department of Aerospace Sciences, School of Engineering

A Computational Aerodynamics Design Group has been established within the Aerospace Sciences Department at Cranfield University, to develop interdisciplinary collaborative activities within the framework of the Cambridge-Cranfield High Performance Computing Facility (CCHPCF) and related eSCIENCE/GRID and Virtual Reality initiatives. In particular the group aims to promote links between the two institutions especially in the areas of CFD and Design Optimisation as part of an Integrated Design Consortium (IDC) involving the Cambridge University CFD Lab and Engineering Design Centre (EDC) specifically. The group is active in three EPSRC High Performance Computing Consortia (HPCC): UK Turbulence (from Direct Numerical Simulation (DNS) to Modelling; leading the subgroup on Applications, Analysis & Control); Combustion (DNS to Large Eddy simulation (LES)); and Applied Aerodynamics (Harrier aircraft descent simulation), as well as GRID applications via the Cambridge Regional eSCIENCE Centre, and the European Research Community on Flow Turbulence & Combustion (managing Special Interest Group 10 on Transition Modelling). The focus is on using the National HPCx and local Supercomputer and cluster resources to perform both fundamental simulations, which provide input to practical modelling, and to utilise the latter for real complex engineering flow applications; thereby helping to replace the most expensive experimental testing and incorporate appropriate Computational Fluid Dynamics (CFD) as part of a more versatile overall design process. With access to all available nodes (many hundreds of processors) on the CCHPCF, local Beowulf cluster with 48 processors and two smaller grid-aware clusters through the eScience initiatives; simulations with tens of million grid points can be undertaken using massively parallel solutions that employ the full range of CFD modeling techniques applicable to industrial applications. CFD is by far the largest user of highperformance computing (HPC) in engineering. The main scientific challenge is the need to gain greater understanding of turbulence, and its consequences for transfer of momentum, heat and mass in engineering applications, including aerodynamics, industrial flows and combustion systems. Availability of HPC has led to significant advances in DNS of turbulence and turbulent combustion, and has encouraged the development of LES for engineering flows. The statistical data generated by DNS have provided valuable insight into the physics of many turbulent flows and have led to rapid improvements in turbulence and combustion modelling for industry. Turbulence is a notoriously difficult subject since analytical solutions to even the simplest turbulent flows do not exist. This results largely from the mixture of chaos and order, and the wide range of length and time scales, that turbulent flows possess. Nonetheless it is essential to continually improve understanding and predictive capabilities, since the overwhelming majority of practical fluid flow problems involve turbulence. Full resolution of the turbulent flow field in most industrial devices remains impractical, and the standard approach is to make use of Reynolds averaging. For statistically stationary flow fields a time average over a period much longer than the correlation time of the turbulence is sufficient and yields a modified set of governing equations known as the Reynolds-Averaged NavierStokes (RANS) equations. The averaging process removes all small-scale phenomena below a mean flow length scale, but the RANS equation set is unclosed due to nonlinearity of the convection terms. A closure model is then required to represent the information that has been lost. Development of closure models for the RANS approach has occupied many researchers for about 30 years. The standard was set quite early by the so called k- model and the rather more sophisticated Reynolds stress model. The strengths and weaknesses of these models are by now well known, and a great deal of research has been devoted to improvement and extension as well as for the proposal of different equation variants. An alternative approach which is now gaining in popularity due to increased computer power is LES. As the name implies, LES attempts to resolve explicitly the large scale flow features, leaving the small scale features unresolved. A formal spatial filtering operation is applied to the Navier-Stokes equations, using a filter width that can be resolved on an affordable computational grid. Once again, unclosed terms arise due to non-linearity of the Navier-Stokes equations, and a closure model is required to represent lost information at the sub grid scale. In this context, the principal advantage of the LES approach over the RANS approach is that most of the turbulent energy containing motions in principle can be explicitly resolved. The main disadvantage of LES is that the computational cost is much greater than for RANS, since the simulation is necessarily three dimensional and time dependent owing to the nature of turbulence, requires finer meshes, and it is no longer possible to take advantage of statistical symmetries or stationarity. Thus, for industrial purposes, RANS remains the more popular choice. The impact of HPC on the field of CFD has been felt in several ways. It has become possible to carry out RANS simulations of much more complex problems involving unsteadiness, coupled physics and complex geometry. LES has become feasible for simple industrial problems and is under active development for application to more realistic geometries. Nevertheless, both approaches are being held back by the

Figure 7: Velocity contours showing the effect of the wheel sting on the near-wake flow-field.

inspection of the circumferential static pressure coefficient, Cp, on the centre-line of the tyre surface, Figure 6 (previous page). Although surface Cp was not measured experimentally, comparison can be made with the work of Hinson [5]. This presented the pressure coefficients on the surface of a Formula One wheel, measured using transducers mounted within. Comparison was made with results taken from tests at the same Reynolds number as, and using a geometrically similar wheel to, this investigation. The results show good correlation and illustrate several important features. CFD predicted the separation 22° late and the base pressure was under-predicted just downstream of separation. Both factors would lead to an under-prediction of drag coefficient and are believed to be responsible, along with experimental errors, for the discrepancy seen in this study.

Support Sting Effects

Conclusions

In this article we have illustrated how the combined use of computational fluid dynamics and experimentation has been applied to three particular fluid dynamics problems: transonic cavity flows, highspeed turbulent jet flow and open-wheeled race car aerodynamics. In each case, knowledge gathered from one analysis technique has been used to assist in the application of the second technique thereby enabling greater under-standing of the flow physics being studied than would have been possible through the isolated use of one or other methodology.

ACKNOWLEGEMENTS

Once the CFD model had been validated, it could be used to determine the effect of the support sting on the flow in proximity to the wheel. The solver settings remained constant, thus improving comparison. Velocity contours behind the wheel reveal a change to the wake vortex structure, Figure 7. Inspection of the remaining data showed that removal of the support sting resulted, for the Champcar wheel, in: · a wheel drag reduction of 2% · an increase in wheel lift of 16% · a reduction in mass flow-rate through the wheel of 83% · a delay of separation by 4°, on the wheel centreline.

Figure 8: CFD prediction of flow entrainment through the wheel hub, from left to right, and into the wake

The authors would like to acknowledge the work of Simon Ritchie and Robin Knowles in support of this paper and input from Dr Nick Lawson. REFERENCES [1] Lawson, N J, Coupland, J M, and Halliwell, N A, `A Generalised Optimisation Method for Double Pulsed Particle Image Velocimetry', Optics and Lasers in Engineering, Vol 27, No. 6, August 1997, Pp 637­656. [2] Wong, R Y T, Enhancement of Supersonic Jet Mixing, PhD thesis, Cranfield University, 2000. [3] Fackrell, J. E., The Aerodynamic Characteristics of an Isolated Wheel Rotating in Contact with the Ground, Ph D thesis, Imperial College of Science and Technology, London, 1972. [4] Knowles, R D, Monoposto Racecar Wheel Aerodynamics: Investigation of Near-wake Structure and Support-sting Interference, PhD thesis, Cranfield University, 2005. [5] Hinson, M, Measurement of the Lift Produced by an Isolated, Rotating Formula One Wheel Using a New Pressure Measurement System, MSc thesis, Cranfield University, 1999.

The slight drag reduction appears to correlate with the later separation, while the additional upper vortex agrees with

28

Aerogram Bringing News from the Horizon

June 2006 Volume 11 Issue 6

29

CFD in Aerospace Design...(continued)

HPCC ­ see example simulation provided by Figure 3. Our research aim now is to use our existing CFD and HPC capabilities to build a working grid-enabled computing environment that will allow us to create and maintain a virtual gas turbine engine in terms of its aero thermal and structural properties, to obtain a numerical solution to flow through the engine by applying advanced HPC techniques, and to enable web browser accessed visualisation of the results. Figure 4 shows the various components of a gas turbine with contributions from our collaborators at Cambridge University and Imperial College, London. The need to model an entire gas turbine, the so called virtual engine, arises from the strong interactions between the various parts of an engine. These interactions couple components in such a way that future optimisation can no longer consider components in isolation. For instance, engine capability, efficiency and life can be significantly affected by icing conditions; the modelling of which is currently a decoupled process. This approach causes significant inaccuracies since the accumulation of ice changes the geometry and thus the airflow. Therefore, icing calculations must be fully coupled to the airflow, thermal and ingestion processes so that integration and data exchange are central to the accuracy of the overall virtual engine. Similarly it is vital to ensure that the engine combustor meets its design objectives in terms of efficiency, aerodynamic performance and minimising the emission of pollutants, without incurring operating limitations due to combustion instability. It is common currently for the entire combustion system to be modelled using a 1D flow network, and this approach has significant advantages of simplicity and very rapid execution times. Nevertheless such models cannot describe the combustion process in sufficient detail, or provide design feedback for geometry changes, and allow mutual interaction with the hot turbine. Instead detailed combustor modelling must be conducted using CFD methods of far greater sophistication and computational expense, namely RANS,

LES and DNS. Structural integrity problems arising from excessive vibration due to flutter, forced response, acoustic resonance and rotating stall also cannot be studied, and avoided at the design stage, without considering the whole machine. Multi-bladerow flow interactions, temperature changes, speed changes and combustion processes affect both the excitation and response. Therefore, we aim to demonstrate that a very detailed full 3D model of an aeroengine can be built, by modular integration of existing technologies in the areas of geometry capturing and management, CFD, heat transfer and cooling, blade vibration, aero-elasticity, combustion, icing, noise and design optimisation. As shown in Figure 4, different components of the engine can and will be developed at different sites using a variety of numerical tools. Vital to this whole project is GRID access to enable each site to exchange ideas and information. At any given point in the design cycle, we propose to have updated hierarchical models from full 3D right down to the overall thermodynamic cycle level. To produce a virtual engine, many components will have to be simulated using different codes, compiled using different compilers and producing data in different formats. Therefore, data handling and connectivity are key collaborative issues to ensure each component is represented as part of the whole engine. The sheer size of this problem as a whole is beyond most pc based cluster systems. However, using the grid enables each component to be simulated on its own, remotely located, dedicated compute resource, and such a distributed design methodology across heterogeneous systems will thus enable many designers to interact with each other. Work is also underway to develop the necessary tools to perform CFD-in-the loop design optimisation. Thus the latest relevant CFD subroutines are being incorporated in a highly modular signposting for design tool, developed within the iDc. This will be extended by the incorporation of cost modelling modules, based on work of colleagues within the School of Applied Science, and the whole software system distributed across the Cambridge-Cranfield grid clusters, via a newly established Virtual Private Network (VPN). The aim is to demonstrate the

June 2006 Volume 11 Issue 6

Figure 1: DNS of developing flame kernel (Dr K W Jenkins).

Figure 2: Representative oil rig module explosion safety case study (Dr K W Jenkins, Prof A M Savill & Dr R S Cant, Cambridge U).

Figure 3: VSTOL hot gas ingestion rig simulation (Prof A M Savill; picture courtesy of Dr G A Richardson, Cambridge CFD Lab).

need for better models of small scale turbulence and related effects, and by the computational costs of very large simulations. Industrial users require rapid turn around and high throughput in order to integrate advanced CFD into the design cycle. Thus the greatest and most immediate impact of HPC on CFD has been the advent of DNS, in which all flow features are explicitly resolved and no modelling is required. DNS has resulted in the development of a new generation of discretisation schemes and solution algorithms with emphasis on high accuracy and high resolution. Data derived from DNS has been used extensively to calibrate existing turbulence models and hence improve both RANS and LES approaches. DNS has been particularly useful in combustion, where turbulent flame interaction modelling has been greatly strengthened as a direct result of available DNS data. However, DNS of turbulent flames is extremely demanding on computer resource requirements, since no turbulence model is used and therefore all scales of turbulence down to the Kolmogorov scales must be fully resolved in addition to the chemistry modelling. In order to deal with this, DNS requires spatial and temporal discretisations which must be both accurate and computationally efficient. Also the use of massively parallel supercomputers is essential in order to simulate physically relevant problem sizes and hence generate useful information in the form of datasets and statistics. Considering the various aspects of turbulent combustion, a DNS code called SENGA was developed, originally by Dr R S Cant at Cambridge University and later in parallel by Dr K W Jenkins. This code is currently

being used to investigate the growth of a flame kernel in a turbulent environment, which has practical applications in spark ignition engines and gas turbines. The DNS code SENGA solves a fully compressible reacting flow. This implies the solution of the compressible continuity, momentum, energy, species conservation equation and the thermal and caloric equations of state in three dimensions and time. Chemistry is treated through the use of a single step reaction mechanism governed by the Arrhenius kinetic rate law which defines a single progress variable ranging monotonically from zero to one depending on the thermochemical state of the system. All spatial derivatives are discretised using a tenth order explicit finite differencing scheme, which has the advantage of allowing complex boundary conditions with spectral like accuracy and is well suited to parallel implementation. Time derivatives are treated using a third order Runge-Kutta method and parallel implementation is through the Message Passing Interface (MPI). Figure 1 shows the growth of a flame kernel, initialised in a turbulent environment on a computational grid of 56million grid points. One major advantage of using the DNS/LES/RANS hierarchy is the ability to simulate explosions in real industrial largescale geometries. This was demonstrated by performing a test simulation of an explosion in an oil rig module. Figure 2 shows rendered flame fronts at a particular instant after two simultaneous ignition events in different parts of the module. The module geometry, measuring 25m by 12m by 8m is a large scale CFD and model validation test case, situated at a remote site in northern England where

British Gas plc has carried out explosion experiments in this geometry. The CAD representation of the geometry was imported and meshed for CFD, and the simulation was achieved using unsteady RANS with laminar flamelet combustion modelling. The initial computational mesh is composed of 2.6million tetrahedral cells. Two levels of solution adaptive grid refinement were applied in the region of the flame to ensure adequate local grid resolution. Many results have been extracted from this large simulation, including overpressures on selected structural members, locations for pressure release vents and insight into practical mitigation strategies. Although the results are taken from unsteady RANS, DNS data was used to validate the combustion modelling employed in this and to pave the way for subsequent LES modelling of this and other real complex geometries. The same approach can be carried over into simulating different structures such as aircraft and space vehicles. The oil rig (Fig. 2) module safety case study is thus just one example of a numerical virtual test bed. Another is provided by our work with Cambridge CFD laboratory on prediction of the descent phase of a Harrier Vertical Short Take-off & Landing (VSTOL) aircraft to assess the possible hot-gas ingestion scenarios. This work is being carried out using up to 30million cell unsteady RANS to LES computations with the latest hybrid adaptive mesh techniques as part of the Defence and Aerospace Research Partnership (funded by EPSRC, DTI and DSTL) on Unsteady Methods for Aerodynamics using HPCx resources from the UK Applied Aerodynamics

Figure 4: Whole engine components computed from left by Cranfield (Prof C P Thompson), Imperial College (Prof M Imregun), Cambridge (CFD Lab team with Prof Imregun.

Figure 5: CFD for Signposting Design Tool (Prof J Clarkson & Dr J P Jarrett EDC).

30

Aerogram Bringing News from the Horizon

31

CFD in Aerospace Design...(continued)

News The CURS of K-LOB

Russell Strand, Cranfield University Rocket Society President

The braving of Bognor

A human-powered flying machine designed and built by Cranfield's Preliminary Year Aerospace Vehicle Design students again featured large among the crazy flying contraptions entered for the 2005 `Bognor Birdman' competition. The Cranfield Bi-plane was launched off Bognor Regis's 15ft pier in the hope of staying airborne long enough to break the 100m flight barrier and scoop the £25,000 jackpot.

The Cranfield University Rocket Society, CURS, is a multinational group of Cranfield students whose aim is to build and fly rockets, chase UK student and amateur altitude records, and gain experience in practical rocket engineering. The first cohort boasted the successful completion and test launch of the `Black Sparrow,' a 2.8m long, 6-inch diameter high-power rocket which test fired to 2,000ft on a solid motor at Cranfield Airport in September 2004. Since then, members have built a second rocket, the `Jenny Wren', and, having had several successful launches with both rockets, are the only students to use hybrid motor technology. One of CURS' aims has always been to take the UK Student Altitude Record ­ currently loitering around the 8,000ft mark. In September we attempted just that at the K-LOB annual launch event. One of the larger exhibitions of the calendar, we saw it as an excellent chance to get our record and some much-desired respect. Our plan was to launch the `Black Sparrow' on an L550 Hypertek Hybrid to some 9,000ft. Not only was this to be a record breaker in terms of altitude, but it would also be the first UK launch of a Hypertek L-Class Hybrid Motor. Some hours after arrival the `Sparrow' was prepped for flight ­ armed with a 75mm diameter, L550 Hypertek Hybrid loaned by Uncle Bob's Rocket Shop and 14 grams of Black Powder for the parachute ejection charges. The beautiful shape stood proud on Uncle Bob's launcher, surrounded by a very large NoX tank and the people holding the step ladder on which I was perched, preparing to arm the flight computers and the pyro-ejection charges. Cranfield's Dr Jenny Kingston and her partner, Howard, were keeping me steady, while Yura Sevcenco was holding the camera nacelle and its nylon fixing bolts. I threw all four arming switches, connected the camera; just four bolts to go and we were ready to leave the pad and start the ignition sequence... I leant over to reach a bolt from Yura's hand, when BANG!

Figure 6: Predicted Pareto trade-off front and compromise optimum blade design ­ only lowest blockage case shown (Prof A M Savill; picture courtesy of T Kipouros & D Jaeggi IDC).

capability of such a system to industry and to research issues such as conflict resolution for simultaneous operation by a team of designers. At the same time a multi-objective Tabu search method is being parallelised and extended to multi-objective design optimisation of generalised 3D turbomachinery compression blading as part of a series of IDC PhD projects. Previous published work from Cambridge (S A Harvey) has shown that single objective optimisation can produce novel blade shapes that rig tests confirm to offer several percentage points performance improvement across a broader operating range. The new studies are demonstrating that further design improvement can be obtained by simultaneously optimising for blade blockage and loss (entropy generated) and then trading off one for the other along the resulting Pareto front. Figure 6 shows the optimised geometry for a compromise point on the Pareto front and the optimization search pattern after 3400 iterations. Finally, to visualise the design cycle, the group has established a state-of-the-art Virtual Reality (VR) Display Facility for VR applications and simulations. This facility is being set up in a cinema-style, purpose built room, within the new extension to the School of Engineering. The VR suite is currently capable of handling an audience of 15 participants, and comprises an approximately 2m by 3m, active stereoscopic projection system, with surround sound, driven directly by local

multiprocessor workstations, and with a Gigabit Ethernet link to the Cranfield end of the CCHPCF, as well as our own local Grid-aware compute clusters. Future developments should include the addition of active tracking for group walkthroughs and a helmet mounted display, combined with a haptic glove, that will allow the wearer to enter within and interact with the simulated environment; as well as a larger 2m by 6m, 100 degree curved screen for fully immersive VR. The Harrier VSTOL simulations are already being output in the required VR format so that these can be viewed as a genuine Virtual Test Bed environment to replace a previous 1/15 scale large industrial experimental facility that proved very costly to operate. Other VR applications should include the 3D analysis of the flame kernel DNS, since deduced model statistics rely on continuity of the flame surface and this will lead on to micro combustor analyses in conjunction with the School of Applied Science; a 3D environment extension to the Knowledge Interface developed by the School of Management; and, as a contribution to Cranfield Health, a virtual heart simulation, (developed again as part of joint project work with Cambridge; funded by EPSRC and MRC) to model the flow and fractionation of electrical currents around the heart wall which can lead to sudden death syndrome - see Figure 7.

In the first half-milli-second, I realised that the ladder was toppling, my supports having all run in opposite directions. I hit the ground just in time to see the nose cone of the `Sparrow' returning from its brief flight, crash through the opening in the ladder and come to rest, peacefully, without killing anyone. Not quite the respect we were looking for; we have, however, since learned that our experience was far from unique and that the BANG was the pyros firing ahead of schedule. Unfortunately, it was too late in the day to re-attempt, so we retired to the sanctuary of our minibus and departed. Later inspection of the avionics bay uncovered a short circuit, something we have ensured cannot happen again. The winter and spring months have been spent improving `Black Sparrow' on many levels, to make the flight safer and more reliable, and also to prevent unfortunate events such as our ordeal at K-LOB. We are negotiating a date for our next launch ­ at the same venue but exclusive to CURS, giving us more attention from some of the experts in the field prior to `pressing the button'. We plan a reattempt with the `Black Sparrow,' and a launch of the `Jenny Wren' on two solid motor flights within the next month. For more information about the society, details of how to sponsor us, a confirmation of our launch date and a report after the event; please visit our website at: www.curs.co.uk or contact Russell at: [email protected]

Having learnt from the previous year's attempt with the `Cranfield Swan', the Bi-plane managed a satisfying 9m, piloted by Robert Morency ­ not bad, considering the Swan only managed to stay airborne for 1.285 seconds and fly 1.4m. Team member Atipu Ponglux said: "It was great to develop and build a design and overcome the various problems along the way ­ educational, but great fun as well." Senior lecturer Phill Stocking said: "Finding an interesting yet challenging project for the students to apply their knowledge to can be a difficult task, but this seemed to fit the bill perfectly. "The students are approaching the competition as though it were a typical aerospace vehicle design project. It's a great hands-on opportunity for them to incorporate their knowledge in areas including structural design, aerodynamics and propulsion."

Figure 7. Mesh Geometry of heart model from physical data provided via Papworth Hospital (Prof A M Savill; picture courtesy of Dr S Rea and Dr P C Dhanasekaran ­ Cambridge CFD lab)

32

Aerogram Bringing News from the Horizon

June 2006 Volume 11 Issue 6

33

News

(continued)

Short courses

Cranfield campus

Our loss is Boeing's gain

Professor Fariba Alamdari, former Head of the Department of Air Transport and Cranfield alumnus, has moved on to pastures new and is now Vice-President of Future Markets at Boeing. Only the second non-US person to be appointed to the board, her appointment is testimony to the close and successful relationships Cranfield has with major industry players. A key part of her role will be to ensure the best strategic decisions are made regarding Boeing's future products, while ensuring a constructive relationship between Boeing Commercial Aircraft and the airline industry.

Memories are made of this

Aircraft Conceptual Design

19 -23 June 2006

Aircraft Performance for Aerospace Managers

4-8 September 2006

This course aims to give the beginner an insight into the procedures and multidisciplinary interactions involved in the process.

Airline Fleet Planning

19-23 June 2006

The purpose of this course is to demonstrate the main points of aircraft performance with the aerospace industry manager in mind. Thus its scope will relate primarily to turbine-powered, subsonic transport aircraft.

In this unique course, you are given a series of presentations by both academic and industrial experts in the field. You will learn how to structure the fleet planning process and how to appreciate and analyse competing and conflicting proposals.

CAP642 Airside Integrated Safety Management JAR-OPS 1 Commercial Air Transport (Aeroplanes) Integrated Safety Management for Airport Ground Operations

4-8 September 2006

Integrated Computer Aided Design (CAD) for Aerospace Engineering

26 ­ 30 June 2006

As a result of attending this five-day course and successfully completing an assignment related to the course contents, you will have gained sufficient knowledge to receive a Certificate in Airside Safety Management.

Memories of Cranfield in October 1946 were rekindled when 13 of the first 100 students held a reunion on the Cranfield campus. Exchanging stories and reminiscing over fond memories, the group was treated to a tour of some of SOE's facilities ­ seeing just how much Cranfield has evolved over the intervening years. The group and their hosts in front of the flight simulator in SOE. Back row (l-r): Prof John Fielding, SOE; Prof Minoo Patel, SOE; Ralph Spenser Hooper; Vic Rogers; Doug Parish; Peter Crawley; Les Morley; Vice-Chancellor Prof Frank Hartley Front row (l-r): Leslie Ellerd-Styles; Arthur Willitt; Peter Wroe; Tony Monk; Bill Morris; John Spillman; Harold Caplan.

This one week short course provides an introduction to the use of integrated computer aided design (CAD) techniques in the aerospace industry.

International Air Law and Regulation Symposium

25-29 September 2006

Computational Fluid Dynamics

June - July 2006 June 2007

The symposium is intended for those working in the air transport industry who wish to both update and broaden their knowledge of air law and regulation.

International flavour

A series of one and two day professional development courses for practising engineers and managers, seeking to enhance their knowledge about this rapidly maturing technology.

Introduction to Flight Simulation

23 - 27 October 2006

Safety Assessment of Aircraft Systems (In association

LARGE EDDY SIMULATION OF TRANSITIONAL AND TURBULENT FLOWS NUMERICAL, PHYSICAL & IMPLEMENTATION ISSUES Cranfield University with the Civil Aviation Authority)

3 - 7 July 2006 27 Nov - 1 Dec 2006

This course is designed for engineers involved with the development of simulator systems, managers and operators of flight simulation facilities and staff responsible for training and procurement.

6 - 8 November 2006

Large Eddy Simulation (LES) is increasingly an established scientific computing approach for predicting transitional and turbulent flows. The intensive short course will cover in detail a number of LES issues, the application of modern LES techniques in aerodynamics, turbulent mixing and combustion, and present the potential of LES in emerging application areas. It is aimed at a broad audience including research scientists and engineers, postgraduate students and industrial practitioners concerned with the development and/or application of LES in aerospace, mechanical, civil and chemical engineering, atmospheric and oceanographic applications as well as other environmental disciplines. COURSE DIRECTORS

Cranfield University has an international reputation and attracts students from around the world. In 2005 46% of students came from outside the UK, half from outside the EU. There are also substantial links with international institutions and corporations around the world including North and South America, Japan, China and India, the Middle and Far East. One notable development specifically connected with aerospace is the potential involvement of Cranfield in the Dubai Aerospace Enterprise University. Professor John Fielding, Head of Aerospace Engineering in the School of Engineering, is pictured at the AVIC 2 Research Centre during a lecture tour in China headed by the China International Talent Development Centre.

Accident Investigation for Aviation Management (formerly Accident Investigation for Aircrew and This course sets out to discuss the various approaches to the problems of assessing the safety of increasingly complex aircraft Operations Executives)

systems.

30 October - 10 November 2006

Aircraft Fatigue and Damage Tolerance

3­ 7 July 2006

This unique course has been specifically designed to meet the practical need of design and maintenance engineers in designing, analysing and maintaining aircraft.

This course brings together the considerable expertise of investigators, operators and researchers to provide safety professionals and operational staff with the fundamentals behind incident and accident investigation and the roles that will be played by operators, regulators, manufacturers, legal and government investigation agencies.

Flight Data Monitoring for Airlines (FDM) and Flight Operational Quality Assurance (FOQA) In Commercial Aviation (In association with the Civil Aviation

Authority)

17-20 July 2006 19-22 February 2007

Large Eddy Simulation (LES) of Transitional and Turbulent Flows Numerical, Physical and Implementation Issues

6-8 November 2006

Professor Dimitris Drikakis (Cranfield University, UK): is Professor of Fluid Mechanics & Computational Science and the Head of the Aerospace Sciences Department. Dr Fernando Grinstein (Los Alamos National Lab, US): is Technical Staff Member, Theoretical and Computational Physics. For contact details, see back page.

This course will provide you with an advanced appreciation of the technical, operational, management and legal issues surrounding a flight data monitoring (FDM) programme, also referred to as flight operational quality assurance (FOQA).

The course is addressed to a broad audience including research scientists and engineers, postgraduate students and industrial practitioners concerned with the development and/or application of LES in aerospace, mechanical, civil and chemical engineering; atmospheric and oceanographic applications as well as other environmental disciplines.

34

Aerogram Bringing News from the Horizon

June 2006 Volume 11 Issue 6

35

Short courses

Cranfield campus

6-10 November 2006

(continued)

(continued)

Shrivenham campus

Aeronautical Engineering - Introduction

6 Nov-8 December 2006

Air Transport Management Seminar

The seminar programme focuses on providing delegates with an insight into the major contemporary managerial and economic issues affecting the airline industry with the aim of broadening participants' knowledge and understanding of how an airline should be managed.

Airport Commercial Revenue Development (Jointly with Concession Planning International Ltd)

13-16 November 2006

This course is designed as a `top-up' for engineering graduates with limited previous knowledge of aeronautical engineering and is primarily run for the benefit of REME officers an alignment module prior to them commencing a Helicopter Engineering course at Arborfield. As such, it has a strong emphasis on helicopter engineering principles.

Guided Weapon Airframes

22-26 January 2007

This workshop for industry managers combines presentations by leading industry professionals, discussions and a terminal concession planning group project.

The aim of this course is to provide an overview of guided weapon airframe design and analysis techniques.

Guided Weapons Aerodynamics

22-23 January 2007

Reliability Analysis

20 - 24 November 2006

The aim of this course is to introduce engineers to the analysis of data from tests or service records and methods of evaluating systems using examples from real life with the emphasis on methods and on appraising the results of the analysis.

The aim of this course is to provide an introduction to guided weapon aerodynamics with the objective of familiarising you with aerodynamics design and analysis techniques via a mixture of lectures and worked examples.

Guided Weapons Structures

24 January 2007

Integration Engineering

5 - 8 Feb 2007

This course aims to provide solutions and balanced views on the above and a number of related problems associated with Integrating Product Development in a fast changing world.

The aim of this course is to provide an introduction to guided weapon structures with the objective of familiarising you with structural design and materials selection via a mixture of lectures and worked examples.

For further details of the professional development opportunities, please contact the Short Course Office at:

www.cranfield.ac.uk/short Tel: +44 (0) 1234 754176

Guided Weapons Propulsion

25-26 January 2007

For details of postgraduate courses, please see:

www.cranfield.ac.uk/prospectus

The aim of this course is to provide an introduction to guided weapon propulsion with the objective of familiarising the student with the design and analysis of guided weapon propulsion systems via a mixture of lectures, worked examples and case studies.

Military Aircraft Technology

19-23 March 2007

If you would like to receive the Aerogram newsletter, please send your contact details to: Cranfield University Aerospace Building 83 Cranfield University Cranfield Bedfordshire MK43 0AL Or contact us by telephone or email: T: +44(0)1234 750111 x 5124 E: [email protected]

The aim of the course is to provide an overview of the main technological disciplines involved in the design of a modern military aircraft.

Information

Aerogram Newsletter

19 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate

120889


Notice: fwrite(): send of 206 bytes failed with errno=104 Connection reset by peer in /home/readbag.com/web/sphinxapi.php on line 531