Computer based mathematical models describing the aircraft evacuation process have a vital role to play in the design and development of safer aircraft, the implementation of safer and more rigorous certification criteria, in cabin crew training and post-mortem accident investigation. As the risk of personal injury and the costs involved in performing large-scale evacuation experiments for the next generation ultra high capacity aircraft (UHCA) are expected to be high, the development and use of these evacuation modelling tools may become essential if these aircraft are to prove a viable reality. This paper describes the capabilities and limitations of the airEXODUS evacuation model and some attempts at validation, including its successful application to the prediction of a recent certification trial, prior to the actual trial taking place. Also described is a newly defined performance parameter known as OPS that can be used as a measure of evacuation trial optimality. In addition, sample evacuation simulations in the presence of fire atmospheres are described.

An improved method of alignment for the laser doppler anemometer has been developed which greatly enhances its capabilities for the examination of the often very small scale complexities of three-dimensional flows. Unlike the alignment methods normally employed, which are subjective, this method is quantitative, using a pin-hole device to measure laser light intensity. This ‘pin-hole meter’ also enables the geometry of the laser beam configuration to be defined in situ and with greater facility and precision than previous methods. Examples are given which illustrate the enhanced capability of the system, the greatest improvement coming from the ability to operate in the difficult to achieve and sustain, off-axis mode of signal collection. With this mode, the improved procedures enable very small measurement volumes, typically 50 micron diameter, while maintaining high data rates, in measurements taken from distances compatible with accessing the flow from outside the windtunnel.

Conventional wisdom states that coordinate complete experimental response data are required to effect a successful model update. This paper examines the influence measured rotational degrees of freedom have on the behaviour of the frequency response function (FRF) sensitivity updating procedure. Three finite element (FE) model-based case studies are undertaken which suggest that very little improvement is afforded by the inclusion of measured rotations in the updating process.

Error-location strategies in the past have yielded little qualitative information about the optimum selection of updating parameters in the model updating procedure. Two further cases are investigated; these suggest that no amount of coordinate completeness will rectify the debilitating effects of an inadequate p-value choice. A method of selecting the number and location of updating parameters — based on the FRF sensitivity method — is presented. This method is shown to yield promising results.

Results obtained from measurements of the dynamic induced velocity field beneath a model rotor are presented. The collective and cyclic pitch of a four bladed rotor of 1·54 m diameter were excited at frequencies up to 1·5 times rotor shaft speed. Flow measurements were taken using hot wire anemometry probes and a laser doppler anemometer. A range of radial and vertical positions near the rotor disc were investigated.

Analysis of the dynamic induced flow response is conducted in the frequency domain and it is found that there are significant radial and azimuthal variations which depend on the frequency of excitation. It is also observed that a change in the character of the inflow response occurs near and above the shaft rotational frequency and that vertical measuring distance from the rotor significantly affects the measured responses.

Some results for the case of cyclic excitation are given. These show, that contrary to momentum theory predictions, the highest induced velocities in the dynamic case do not occur over the area of the disc where the blade pitch is at its maximum.

Overall, the results show that the dynamic induced velocity field is highly complex and heavily influenced by the distribution of time varying shed vorticity within the wake.

The methods of non-linear dynamics are applied to the longitudinal motion of a vectored thrust aircraft, in particular the behaviour at high angles-of-attack. The model contains analytic non-linear aerodynamic coefficients based on Nasa windtunnel tests on the F-18 high alpha research vehicle (Harv). The equilibrium surfaces are plotted against thrust magnitude and thrust deflection and are used to explain the behaviour. When the aircraft is forced with small thrust deflections whilst in poststall equilibrium, chaotic motion is observed at certain frequencies. At other frequencies, several limiting states coexist, e.g. a chaotic attractor and a limit cycle, or two limit cycles. The steady state behaviour then depends on the initial conditions. The non-linear pitching moments are shown to have a significant effect upon the model. Omitting these terms, the region of chaos shrinks, this is not observed for lift.

The integration inverse method has been extensively investigated in the past decade; however, none of the researches fully addresses the stability analysis of the method that is crucial to successful implementation. This paper presents a stability test to analyse the global stability of the integration inverse method for linear time-invariant systems. A stable solution may be obtained from careful selection of the discretisation interval using the proposed stability test. A discrete model is derived to approximate the Newton's scheme in the inverse method. With this approximate model, the stability of the inverse method can be examined. The stability test is exact for linear systems and can be extended to the inverse method for non-linear aircraft simulations by considering an equivalent linear model for each point of the flight envelopes. Guidelines for selection of appropriate reference points in the inverse simulation are given.