A Study On Rejection Rate Of ESO Method And Multi-objective Structural Optimization Of The Flywheel

With the development of science and technology, more and more requirementsabout performance and economical efficiency have been proposed for the structuralmechanics. The structure optimization design is such a rapid development of the designfield. It makes designer more active and creative. The structural topology optimizationlays the foundations for subsequent shape and size optimization.Fly of rebirth is a kind of important energy storage regulating speed fluctuationstructure of mechanical parts, so this paper makes further research on the flywheeltopology design on the basis of previous studies. Based on the consideration of thestructural stiffness, strength, and requirements of larger moment of inertia and lowermass, multi-objective optimization design of flywheel, this research was carried out onthe flywheel multi-objective optimization design. Also, multi-objective optimizationdesign has great value for engineering application.This paper did the research on the flywheel multi-objective topology optimizationby using the DFP (Davidon Fletcher Powell) method and the hybrid method containingthe hierarchy, multiply-divide, and main target, respectively. Then, a comparison of theresults obtained by these two methods will yield additional insights. For the DFPmethod, the paper constructed the flywheel evolutionary structural optimization methodbased on global standards unit sensitivity, and put forward a new method of sensitivityfiltration unit, and finally got the better optimization result of the flywheel. It provesthat the DFP method has standout efficiency and practicability in dealing with themulti-objective problems. For the hybrid method, the first-layer target took the qualityand inertia of flywheel into consideration by using hierarchical optimization method.The variable thickness flywheel was designed by using multiply-divide method. Thesecond-layer target took the stiffness into consideration by using the main target method.It set the stiffness as the main target and set the intensity of the flywheel as theconstraint condition. On the basis of the first layer variable thickness flywheel ESOmethod, the flywheel is optimized. Finally, it got an excellent optimization result of theflywheel, and proves that this method also has standout efficiency and practicability indealing with the multi-objective problems. In addition, there is randomness for the values for deletion rates and the rate ofevolution in ESO method. Each researcher determines the values according to their ownexperience, and there is no certain mathematical theory. So this paper has a summaryand statistic on the values of the deletion rate and the rate of evolution, which aredetermined by evolutionary structural optimization experts and scholars. And based onthis summary and statistic, the weighting method is used to integrate views of all theexperts to confirm the deletion rate value, intended to disperse and reduce the individualdesigners subjective error on the values, making the values is more scientific andreasonable, but the results are not satisfactory.For an initial full design area, a reasonable strategy for element deletion shouldgradually reduce the number of element deletion as the iteration proceeds, while thestructural stress distribution is more uniform. This paper constructs a more simpledynamic deletion rate based on the current state of stress. The dynamic deletion ratemake the element deletion number gradually reduce as the iteration proceeds, so that theelement deletion rate can be adjusted with the relevant information of the optimization,and the optimization process can be effectively improved. In this method, based on theconcept of standard deviation in mathematical statistics, a new stress uniformity indexof the structure is used, as a judgment of the stress uniformity of the structure. Throughthe analysis of the typical cases, this evolutionary optimization method has theadvantages of rapid convergence, smooth and steady performance characteristics. Andapply this dynamic deletion rate to the flywheel multi-objective structural topologyoptimization.