Please choose your delivery country and your customer group
Fixture design is a highly heuristics dominated area and efforts to automate this operation have been largely focussed on generative design approaches. This paper however makes a fundamental shift to a variant design approach using case based reasoning to automate the fixture design process. The system is based on two key concepts: one, to extract geometric information from a solid model representation of the workpiece and structure the extract into the current 'case' and two, to compare the current 'case' against a 'case base' of past design successes to find the nearest match. After the 'match' has been identified, an adaptation module reconstructs the fixturing system used in the 'match' on the current workpiece and presents it to the operator for verification. Subsequently the newly generated design tagged with the result of the verification is appended to the case base and the case base is reindexed. Key benefits of this approach include the disassociation from the complexities associated with feature recognition and comparison used in traditional systems and the continuous process of knowledge acquisition which betters the system with each pass. The system is also general enough to be able to handle all classes of components and does not require extensive reconfiguration or maintenance when moving from one class of components to another.