Stress Optimization of Bike Frames

David Recasens, Juan Brion (Linkopings University)

CHALLENGE - The objective of the project is to find the best geometry configuration for each bike frame, focusing the search on minimize the weight and the stress suffered. 

SOLUTION - The search of the optimum geometry design was carried out with the aid of modeFRONTIER. By integrating CATIA, modeFRONTIER executed  the optimization by switching the value to the input variables (bike frame parameters) between bounds and allowing monitoring the outputs results, helping to find the best solution reducing the waste of time. The workflow of the program consists of nine inputs (nine bike frame variables) which lead to two outputs (weight and stress through the maximum Von Mises) with minimizing objectives. We used Random configuration at DOE properties with 30 number of designs and NSGA-II with 50 number of generations at the scheduler. As result, there are 1500 evaluated designs. This is due to there is only over 30 % of feasible results at the simulations and in order to have a decent estimation around 500 feasible results are needed. 

BENEFITS - modeFRONTIER allowed  to analyze hundreds of designs, achieving a great precision to select the best bike frames geometries.