Optimal design of flexspline structure based on approximation model

An efficient multi-objective optimization procedure to maximize the fatigue life of a reform flexspline (FS) structure is presented. Improper design of FS structure may cause the malfunctioning of harmonic drives (HD). After a brief introduction of the existing design flaw on the traditional cylinder FS, a new structure with minimal oblique angle in FS cylinder wall is proposed which can improve the meshing condition of gear. The finite element model is employed whose result shows that the fatigue life of HD can be greatly enhanced by changing the cylinder shape FS to oblique angle one. Key structure variables of oblique shape FS are identified through sensitivity analysis which based on response surface method (RSM). Furthermore, the multi-objective algorithm which is basis on the RSM approximate model is employed to search for the key variables pareto optimal solution of FS. Stress and fatigue life are taken as objectives in multi-objective model which are solved by non-dominated sorted genetic algorithm (NSGA-II). The computational result shows that the most suitable oblique angle should be set at alpha = 2.5o which prolongs the fatigue life by 365.2%. A new oblique shape FS structure form is provided to improve the traditional FS structure. The multi-objective optimization method is used to further optimize the variables of the oblique shape FS. The better FS structure can be obtained through this optimization procedure. According to the optimal oblique shape FS, the engineers could design high quality HD to ensure the efficient operation of industrial robots.