@conference {3303, title = {3303. Evolutionary Feature Based Weight Prediction}, booktitle = {62nd Annual Conference, New Haven, Connecticut}, year = {2003}, month = {5/17/03}, pages = {23}, publisher = {Society of Allied Weight Engineers, Inc.}, organization = {Society of Allied Weight Engineers, Inc.}, type = {23. WEIGHT ENGINEERING - STRUCTURAL ESTIMATION}, address = {New Haven, Connecticut}, abstract = {Many weight prediction methods operate ?analytically,? deriving an ideal structural weight, and then accounting for non-optimum structural penalties using a factoring approach. This paper presents an alternative weight prediction philosophy that combines elements of traditional mass accounting with estimated sizing data to produce detailed and accurate weights; the component weight is calculated from a set of detailed parameters using a simplified traditional accounting method. These detailed parameters describe the predicted design of the particular component in its entirety, identifying and sizing all features. The large number of detailed parameters required would normally preclude the use of this method during the early project phases ? where weight prediction is most valuable. However, these detailed parameters can be obtained using their respective driving parameters. A generative design approach allows the prediction of relevant features. Using parametric relationships based upon initial load estimates enables the prediction of detailed sizing. The result is a versatile method which can be implemented at any stage of a project to generate predicted weights. The resulting weights contain correct causality for ideal structural weight and simultaneously integrate weight resulting from manufacturing constraints and other design considerations. Other benefits of this method include being able to identify the true weight drivers of a component and ascertain the weight impact of particular design alterations and manufacturing methods.}, keywords = {23. Weight Engineering - Structural Estimation}, url = {https://www.sawe.org/papers/3303/buy}, author = {Anna Baker and Douglas Smith} }