A Feature-Based Approach to Function Modeling and Reasoning for Thermal-Fluid Systems Engineering

Abstract

This research is focused on extending the language of function modeling by developing CAD-like features that are built using primitive functions. These features are reusable units of functions that represent the working principle of a system of higher level of abstraction than the ones of the primitive functions published in prior literature. This approach can be used to also establish an alternate concept/working principle that satisfies the same features, and thus enabling the integration of function models into previously established concept generation programs. It is expected that designers perceive concepts at different levels of abstractions. As a result, the feature-based approach will be of high significance when designers model a complex system since this approach will allow traversing the abstraction levels of the concept within the same model. The features that recur in complex systems are identified, formalized as vocabulary and grammar within the language, and implemented in a software program with reasoning algorithms using a set protocol. The implementation of this protocol is demonstrated in a particular domain, in this case, thermal-fluid systems. The extension of the protocol is shown by modeling additional devices that lie in the food-processing domain. The feature-based models developed will be used to provide system-level support and reasoning, thereby extending the potential of the language. The change in ease of modeling and model build time is tested using a preliminary test by modeling several devices. The impact of this approach on the consistency of models will be tested with human subject studies.