Computational Methods to Predict and Avoid Design Failure
Conference Paper
Overview
Identity
Additional Document Info
Other
View All
Overview
abstract
Component selection can be a difficult task for designers, and the components they choose can have a large impact on the robustness of the design. Using previous methods to predict and identify potential failure modes, known as the function-failure design method (FFDM), the impact on failure of selecting a particular component over another can be explored based on failure results from previous design endeavors using the same component. This assists designers in selecting the component that is best suited for the application. Since the predicted distribution of failure modes changes depending on the selected component, failure reduction is possible through component selection. Through this method of component selection, risk can be decreases and potential failures can be eliminated. Experiments based on undergraduate student competition design projects are presented to illustrate this methods ability to predict failure modes. Initial results indicate that the predictions are accurate and meaningful to designers. The experiment also serves as initial validation previous work in the area of failure prediction.
