Department of Engineering - Annual Report 1996/97

Energy And Fluid Mechanics

ENERGY
Acoustics
Control of Emissions from IC Engines
Computational Combustion
Stirling Engines
Fluid Mechanics in Process Metallurgy
Magnetohydrodynamics
Nuclear Engineering

FLUID MECHANICS
Fluid Mechanics
Compressible Aerodynamics
Coastal Engineering
Computational Fluid Mechanics

References

Energy

Acoustics

The interaction of fluid motion and sound(A31) continues to be a research speciality even though the subject has, because of its importance to both aeronautical and marine applications, attracted a great deal of attention for almost half a century(A30). The general characteristics of the sound field can be understood in terms of elementary acoustics(A28), but detailed predictions require consideration of the causes of flow unsteadiness. Such unsteadiness often arises from flow instability. A new method for predicting the transition to turbulence on swept aircraft wings has been developed by exploiting ideas from the theory of absolute instability(A65). The concept of absolute instability has also been used to model vortex shedding from blades(A68,A69) in order to understand the excitation of acoustic resonances in compressors. The distortion and evolution of vorticity is a significant source of sound in a variety of aeronautical and marine applications. The ingestion of atmospheric turbulence by the fan of an aeroengine leads to a tonal noise, produced by the repeated chopping of the distorted eddies by the rotated blades, and can be particularly strong in ground tests and in aircraft approach conditions(A49). Consideration of noise generated by the interaction of a blade row with oncoming gusts and wakes is being used to develop an analytically-based prediction scheme for rotor-stator interaction noise in compressors(A29,A57,A58). The scattering of vorticity waves by obstacles in supersonic flow leads to sound(A56). A model based on a superposition of vortical structures has been used to provide physical insight into the noise producing events in the wall region of a turbulent boundary layer(A26). The model not only predicts the high frequency and high spanwise wavenumber range of the surface pressure spectrum well, it also clarifies the contribution of viscous effects to the low wavenumber spectral elements.

The interaction between acoustic waves and combustion can lead to intense self-excited oscillations. Industrial burners and aeroengines designed for low chemical emissions are particularly susceptible. Work in progress aims to understand the instability in lean premixed gas combustors (funded by Rolls-Royce plc) and in liquid fuel-spray atomisers (funded by EPSRC). A nonlinear model of the generic problem of a ducted flame has been developed and used to investigate aspects of active control(A27). The maintenance of quiescent flow conditions by active management remains an ongoing research area. Active controllers usually rely on control systems carefully adapted to the specific job in hand. A control method based on maximising the rate of energy extraction from a wave field(A34,A35) has resulted in a sound and vibration suppressing technique(A36) and an invention patented in Japan.

The interaction of flow with flexible surfaces can lead to surface vibration and flow instability. Collaboration with medical colleagues is concerned with avoiding instability within the human airways. Consideration of the interaction between underwater structures and the surrounding flow has resolved some anomalous wave phenomena(A55).

Control of Emissions from IC Engines

Dr N. Collings

In the last annual report, initial studies on a novel very high frequency response sensor for detecting oxides of nitrogen (NOx) were reported. This work has resulted in the development of a commercial instrument(A60). Further detailed studies continue within the department, aimed at trying to improve our understanding of the device and optimise it. Initial results from the device have revealed a great deal of detail about the operation of automotive catalysts under transient conditions - detail that was hitherto unmeasurable.

In a related project, we have been looking at ways of rapidly changing the mixture strength of a gasoline engine without causing torque fluctuations. If the engine can be operated in this way, there are opportunities for improving the fuel economy of engines, while reducing emissions levels even further.

Following the installation of a new transient engine dynamometer facility, in collaboration with Professor K. Glover of the Control Group (previously reported), three researchers have been appointed. All the work is related in part to emissions control, but the Energy group focus is on catalyst transient behaviour.

Another joint activity, again with Professor Glover, and also with Dr R. Lambert of the Chemistry Department, has just been established, with the objective of achieving improved emission system performance under real-life conditions, especially with regard to long term system degradation.

Computational Combustion

Dr R.S. Cant

Research in computational combustion continues to grow in strength, and is now firmly linked with the newly-formed CFD Laboratory across many areas of common interest.

A full range of combustion modelling capabilities has been established, including Direct Numerical Simulation (DNS), Large-Eddy Simulation (LES) and Reynolds averaged Navier-Stokes simulation (RANS). Work in DNS is being funded by EPSRC through two projects covering the physics and the numerical analysis of the reacting Navier-Stokes equations. DNS poses particular challenges in terms of accuracy and computational efficiency, and these are being met through the use of novel wavelet-based discretisation methods together with highly-efficient parallel algorithms(A14,A15,A16). Two locally-developed DNS codes are in use for incompressible and compressible problems respectively. Computer resource requirements are large and are being met using local resources, including the Cambridge High Performance Computing Facility, as well as by the National facilities such as the Cray T3D and T3E parallel supercomputers at Edinburgh.

Industrially-sponsored research projects are supplementing EPSRC support in the areas of LES and RANS modelling, with significant support from Shell Research, HSE, EGT, Hitachi and Ricardo Consulting Engineers. The range of applications is broad, covering gas turbine combustion, spark-ignition automotive engine combustion and the prevention of accidental explosions in the offshore oil industry(A12). All of the computer codes in use have been developed locally, and include simple structured-grid testbed codes for both compressible and incompressible problems as well as unstructured adaptive codes capable of handling the most complex geometries.

Stirling Engines

Dr A.J. Organ

A text, The Regenerator and the Stirling Engine has been completed, and is with the publisher for publication in November 1997. It bridges the gap between classic regenerator theory, conditions in the Stirling cycle machine, and the general solution of the regenerator and conjugate heat transfer problems pioneered at Cambridge(A50, A51). An experiment has been devised which reveals the character of Stirling engine pumping losses. Results fully vindicate the predictions of the analytical work and computer simulation hotly disputed elsewhere. An explicit and largely analytical method has been developed for striking an optimum balance between pumping loss and heat transferred in the conventional steady-flow exchange.

Fluid Mechanics in Process Metallurgy

Dr P.A. Davidson

Fluid mechanics plays a critical rôle in the production and casting of metals. For example, aluminium is produced by electrolysis in large reduction cells. This process is highly energy intensive, requiring around 30 MWatt hrs to produce one tonne of semi-finished product (about seven times the energy required to produce steel). Yet the bulk of the energy expended in reduction cells is not required for electrolysis, but rather is lost in resistive heating of the electrolyte. Attempts to increase the efficiency of these cells, by reducing the volume of electrolyte, have been fundamentally limited by the appearance of an instability which manifests itself as a sloshing of the electrolyte and is directly triggered by the reduction in electrolyte volume. The fundamental nature of these instabilities is just beginning to be understood(A22,A23,A43).

In the casting industry, on the other hand, fluid flow plays a critical rôle in determining the metallurgical structure, quality and surface finish of an ingot. There are vigorous turbulent flows within the liquid zone of a partially solidified ingot, driven by thermal buoyancy, by compositional variations throughout the melt, and by the flow of liquid into the mould. This flow is the dominant mechanism for transporting heat, chemical species and crystal fragments within the liquid metal pool. Modest changes in the structure of this flow can produce a large difference in the distribution of impurities and dendrite size and morphology within the final ingot.

Research Council awards in this area are supporting our work on: instabilities in aluminium reduction cells(A22,A23, A43);the influence of fluid flow on macrosegregation in aluminium ingots(A23); electromagnetic control of liquid metal in twin-roll casting(A67); the control of high-temperature titanium jets in spray-forming processes; and the fluid mechanics of vacuum-arc remelting.

Magnetohydrodynamics

Magnetic fields can be used to levitate, pump and stir liquid metals(A21), to suppress natural convection(A2,A3,A4,A32), and to reorganise the structure of turbulent flows(A5,A6,A7,A8,A19,A20,A48). We are actively involved in all of these areas. In MHD turbulence we are examining the rôle of magnetic fields in preferentially suppressing transverse components of angular momentum. This leads to a strongly anisotropic turbulent structure, reminiscent of turbulence in stratified or rotating media. This may be quantified using a variant of the classic Kolmogorov/Landau model for the decay of conventional, isotropic turbulence(A19,A20). We are also examining the behaviour of quasi-two-dimensional turbulence, induced by very intense magnetic fields(A5,A6,A7,A8). The tendency of the flow to develop large 2D eddies was observed as well as the expected poor heat transfer predicted by maximum entropy theory.

Regarding buoyancy-driven flows under a static magnetic field, asymptotic models have been developed(A3,A18) and have led to an understanding of the conditions for the damping of the flow by the magnetic field. This has practical importance in the fields of crystal growth and metallurgy. Moreover, magnetic damping provides a cheaper alternative to many current micro-gravity experiments. We have started a collaboration with two French laboratories (Madylam and CEA-CEREM) to use magnetic damping to help measure the value of solute diffusivities, which are otherwise extremely sensitive to a parasitic convection(A2, A32). Such precise values are crucial to model the growth of semi-conductors used in opto-electronics.

Nuclear Engineering

Development continues(A53,A54) in the optimisation schemes for light water reactor in-core fuel cycles supported by Nuclear Electric (British Energy). Studies in transport theory have led to an assessment of the fluctuations of predicted eigenvalues in criticality studies(A9). Work with compound systems having constant mean specific failure and repair rates has been completed and studies initiated of systems with ageing parameters(A17). Consideration of heat transfer and thermodynamics has led to a novel interpretation of the thermodynamics of surfaces. Work has been started in optimal design theory. A study of reactor stability driven by Xenon oscillation was published(A42) and work on the conservation of momentum of charged particles in magnetic fields was completed.

The monograph Finite Element Methods for Particle Transfer was published together with Volumes 24 and 25 of Advances in Nuclear Science to Technology(A40,A41). An historical article was completed(A39). A lecture on managing the law(A38) was edited and published.

Fluid Mechanics

Fluid Mechanics

Professor W.N. Dawes

Experimental work this year has been much disrupted by major re-building work within the area laboratory aimed at rationalising our facilities and regrouping with the focus very clearly on our supersonic blow-down tunnels and large low-speed tunnel. These two facilities are enormously valuable to us and much in advance of those available these days to University groups elsewhere. We are in a strong position to respond to the return to fashion of experimental work in its own right (rather than as just an adjunct to CFD).

Compressible Aerodynamics

Dr H. Babinsky

The investigation of passive control of shock/boundary-layer interactions has been completed(A33). Findings indicated a possible reduction of wave drag for transonic wings, however an increase in viscous drag was also observed. An improved mechanism for controlling shock/boundary-layer interactions is being investigated for swept and unswept shock waves as part of the EUROSHOCK II programme. It is hoped that active control incorporating suction can avoid the increase of viscous drag while retaining the beneficial effects on wave drag. For this purpose one of the supersonic wind tunnels has been modified to incorporate an ejector-based suction system. First tests with the new device have indicated that the performance meets the specifications.

A paper on earlier work concerning roughness effects in shock/boundary-layer interactions has now been published(A10).

Coastal Engineering

Dr J.F.A. Sleath

Work has continued on wave-induced sediment transport. Preliminary results of an experimental study of the movement of sediment under extreme conditions have been submitted for publication. Theoretical limits on the formation of plugs in oscillatory flows have been investigated.

Velocity measurements have been made close to rippled beds in oscillatory flows. These measurement shave been compared with numerical solutions for the flow. A numerical and experimental study has also been made of the effect of bed roughness on wave-induced drift. This is an important problem for anyone concerned with the transport of pollutants by wave action.

Finally, the effect of sediment movement on the turbulence intensity and on bed friction in oscillatory flow has been investigated and the results have been submitted for publication.

Computational Fluid Mechanics

Professor W.N. Dawes
Professor A.P. Dowling
Dr R.E. Britter
Dr R.S. Cant
Dr A.M. Savill

This year has seen the consolidation of the opportunities presented by the newly completed CFD laboratory. We are now building on our core areas of competence: unsteady, solution-adaptive flow simulation in complex geometries(A24); combustion modelling(A62) (including DNS/LES); transition/turbulence modelling (again including DNS/LES)(A15,A16,A66); and acoustic/combustion coupled instabilities. These core research areas are supported by a series of application-based contracts with industry (UK and abroad), HSE, EPSRC, CEC, etc. looking at real-world problems including oil rigs(A12), gas turbine combustion chambers(A25), and airfoils with riblets(A59). A number of publications are starting to appear. Future work will include more on explosion mitigation, dispersion (of flammable fuel, pollution, etc.) and the use of DNS/LES to improve industrial-level turbulence models.

[Table of Contents]

Last modified: 8 August 1998