Aircraft system simulation methods in recent Airbus programmes

摘要

The A380 programme was for AIRBUS an opportunity to establish new methods for system modelling. Compared to previous programmes, the aircraft has a larger number of system controllers and a greatly increased level of data exchange in the aircraft network. At the same time AIRBUS required their equipment suppliers to provide simulation models. This policy was intended to allow more complete models to be available on all types of simulators, increasing the overall fidelity. This paper presents the approach that was taken to improve the system modelling phase in order to face these technological changes. In this context, the following issues are discussed: 1. Modelling phase: How the fidelity of models was improved by using formal specifications of the system controllers from the AIRBUS Design Office. Development of the models with particular attention paid to respecting the aircraft equipment interface. Extension of the use of modelling tools based on logic diagrams (such as Scade or Matlab Simulink) to reduce computer programming errors and to increase the efficiency of modelling. 2. Verification phase: Use of automatic stand-alone verification and non-regression checking of the model formal specifications to consolidate the development phase and to carry out in depth checks prior to simulator integration sessions. 3. Validation phase: The same models can be used on the different A380 simulators, from the desktop simulator to the integration simulator (known as Aircraft 0), providing a high level of model maturity and fidelity very early in the Aircraft Programme (by interfacing them with real system controllers and making direct comparisons available), and before making them available for use on training simulators. The future trends of simulation solutions are also discussed, in particular the way model elements developed for the A380 simulators could be reused on forthcoming A400M and A350 simulators. Examples from the A380 Electrical and Engine simulations are used to emphasize the implementation of these methods and the levels of fidelity and maturity these models are expected to achieve.