AbstractThis thesis attempts to answer the question Can we devise a language for interpretation of behavioural simulation of engineered systems (of arbitrary complexity) in terms of the systems’ purpose? It does so by presenting a language that represents a device’s function as achieving some purpose if the device is in a state that is intended to trigger the function and the function’s expected effect is present. While most work in the qualitative and model-based reasoning community has been concerned with simulation, this language is presented as a basis for interpreting the results of the simulation of a system, enabling these results to be expressed in terms of the system’s purpose. This, in turn, enables the automatic production of draft design analysis reports using model-based analysis of the subject system. The increasing behavioural complexity of modern systems (resulting from the increasing use of microprocessors and software) has led to a need to interpret the results of simulation in cases beyond the capabilities of earlier functional modelling languages. The present work is concerned with such cases and presents a functional modelling language that enables these complex systems to be analysed. Specifically, the language presented herein allows functional description and interpretation of the following.
• Cases where it is desired to distinguish between partial and complete failure of a function
• Systems whose functionality depends on achieving a sequence of intermittent effects
• Cases where a function being achieved in an untimely manner (typically late) needs to be distinguished from a function failing completely
• Systems with functions (such as warning functions) that depend upon the state of some other system function.
This offers significant increases both in the range of systems and of design analysis tasks for which the language can be used, compared to earlier work.
|Date of Award||04 May 2006|
|Supervisor||Neal Snooke (Supervisor) & Chris Price (Supervisor)|
- functional representation
- design analysis