Dependency-based collaborative design: a comparison of modeling methods

摘要

The ability to provide support to a group of designers, analysts and other users who are collaborating on anrnevolving design requires the dual capabilities of managing options for an individual designer while at the samerntime managing the dependencies between different sets of designer’s options. For example, if designer A isrncreating a design for a helicopters hydraulics system and this is dependent on electrical power (EP) from arnsub-system from designer B then how can the workflows, decisions and options of the two designers bernmanaged so that each can understand the implications of their own design decisions and more importantlyrnthe implications and design decisions they force on others? The proposed CAPS system employs two dependencyrnreasoning engines: one handles quantitative values and other qualitative ones. The quantitative enginerncan identify that a motor with an output of 3000 rpm allows a generator to output 100V or that a hydraulicrnpump is dependent on the 240v output of the motor. Alternatively, the qualitative engine could rule out severalrnmotors options if the overall design state that the weight of a helicopter’s transmission needs to be comparablernto that of the engine or the positioning of a sensor makes it susceptible to an EM process that could affect itsrnfunction. A mapping capability is provided allowing analysis to be passed between the two engines. Three differentrntypes of dependency-based quantitative engine designs have been developed and evaluated. The firstrnquantitative engine design focuses on modeling components at their output level and propagating level valuesrnfrom component to component, component to sub-system, etc. The second quantitative engine design focusesrnon modeling components at the node level so as to identify key components, sub-systems, etc. in terms of theirrnoverall dependency to the design. The third engine employs a hybrid of the previous two approaches and wasrnidentified through feedback from designers. These engines are designed to be complimentary with the nodebasedrnand hybrid approaches being used to identify the key capabilities and dependencies of the design.rnThe output-based approach is then used to explore in greater detail the outputs and dependencies of therncomponents and sub-systems identified as key via the node/hybrid-based approach. The proposed CAPSrnarchitecture has been initially evaluated against a large collaborative design task involving the design of a helicopter’srnelectrical, hydraulic, structural and mechanical systems.