Evolution of model-based design in aerospace

31 March 2011

Model-Based Design continues to grow within the aerospace and defence industries. Matt Behr examines the reasons behind that growth and explores future trends in its adoption. Particular focus is given to its use in the development of certified systems and on large multi-organisation programmes

The development of aerospace and defence systems presents unique challenges. The first challenge is managing their extraordinary scale and complexity. Frequently, these projects are ‘systems of systems’, requiring integration of disparate dedicated systems. Next, low production volume means that nonrecurring engineering costs are carefully scrutinised. One-time costs for research, design, and development cannot be distributed over thousands or millions of units. Lastly, testing these systems can be difficult, costly, and unsafe.

For example, commercial and military satellites cannot be fully tested on the ground and conducting flight tests on new aircraft is both expensive and hazardous.

Aerospace and defence organisations have long utilised modelling and simulation to address these challenges.

Simulation technologies, including commercial tools such as Simulink, have evolved to support engineers throughout the design, development, and testing cycles.

Early in the design cycle, simulations are used to understand and analyse system behaviour. As the functional and performance requirements of systems have evolved, so too have simulation and analysis capabilities. Many organisations still use custom FORTRAN based models in their design processes.

These custom environments, while effective for their original task, can be difficult platforms on which to add modelling capabilities. This dynamic has led the industry to turn to commercial-off-the-shelf (COTS) simulation packages. An example of this evolution is the addition of discrete event simulation to Simulink. NASA and TriVector Services recently used these capabilities to analyse the impact of communication latencies on the Ares I rocket.

In addition to providing design insights and facilitating verification via simulation, modelling tools supporting code-generation allow models to be re-used throughout the project life cycle.

Indeed, re-use is one of the key advantages of Model-Based Design. Code generated from models is often used in real-time hardware-in-the-loop testing. By running models in real-time with hardware I/O, engineers can compare the behaviour of the processors and hardware with the behaviour of the simulated components.

Code generation also enables organisations to re-use algorithmic models in production systems.

Production code is generated automatically, rather than reimplemented by hand, saving time and eliminating errors.

Having benefited from its utility in simulation, verification, and production implementation, organisations are now looking to solve additional challenges using Model-Based Design. Specifically, organisations are attempting to ease the burden of compliance with industry standards and enable integration testing via simulation on multi-organisation programs.

High-integrity programs requiring compliance with industry standards such as DO-178B present unique challenges. The increased burdens of testing and artifact generation significantly increase cost.

Model-Based Design helps engineers achieve certification to safety standards by supporting requirements traceability, verification, and documentation. These capabilities span multiple design stages.

For example, requirements linked to model are inserted as comments in generated code. Qualification kits, available for several verification tools, can reduce the amount of manual review needed.

It is also increasingly common for organisations to adopt Model-Based Design on large programs spanning multiple organisations. This allows system-level performance to be assessed and integration issues to be uncovered much earlier in the design process.

When detailed models from multiple organisations are combined, resulting models can contain hundreds of thousands of blocks. Modelling tools have evolved to meet these challenges with improved support for large-scale modelling, including support for composite models from other model files and support for signal buses.

Modelling standards are also becoming important for these multi-organisation programs. Much like coding standards were adopted to facilitate team development and sharing of source code, modelling standards are being developed to support collaboration at the model level.

For example, the Orion Guidance, Navigation, and Control (GN&C) MATLAB and Simulink Standards document describes the modelling standards and guidelines that the Orion Crew Exploration Vehicle Flight Dynamics Team used for GN&C algorithm development.

The standards provide guidelines for several aspects of the GN&C models, including stylistic rules, modelling tool selection, and configuration settings, which affect model readability as well as the generated code.

As Model-Based Design continues to evolve, it is enabling a diverse and expanding group of leading organisations to improve efficiency, increase re-use, and meet new challenges in developing modern aerospace and defence systems.

The author is aerospace and defence industry marketing manager, MathWorks

Contact Details and Archive...

Print this page | E-mail this page