Gear Optimization Of Power Split Device For Hybrid Electric Vehicle

Power split device (PSD) can be considered as an evolution of traditional gearbox, it canwell solve the problem of power coupling and energy assign. Compared to the traditionalgearbox, PSD can satisfy the requirement of torque and rotational speed decoupling controlwhen the hybrid electric vehicle is continuous operating with different loads, and better fueleconomy may be realized through control engine works based on the optimal fuel consumptioncurves and in stable state.This paper takes gears mesh of PSD as research objects to analyze the influence fromaxial deflection and axial backoff, and how to optimization of PSD gears so as to realize driveefficiency and lightening design based on the project of National Natural Science Foundation ofChina: Design Theory and Method for Power Coupler of Hybrid Electric Vehicle in Multi-fieldCoupling (51075179). The research is done in view of the following aspects:First of all, the influence on the gear mesh stress from axial deflection and axial backoffwas analyzed by simulation, and the results show that the phenomenon of gear mesh stress andstress concentration on the transverse tooth profile were obviously increased. Then thethermal-structural coupling analysis has been done, the friction dissipation increases markedlyas the stress increases.Secondly, the modulus, tooth number of planet gears, transmission ratio, axial deflectionand axial backoff of planet gears were taken as input variable, a group of30sample points andanother group of10sample points were selected by the design of experiment method ofoptional Latin Hypercube. The responses include highest contact stress on the tooth face and thefriction dissipation, were calculated by heat-structure coupling analysis through Abaqus. Thenthe responses were collected and used to establish agent model separately using polynomialresponse surface models, radial basis function, and Kriging model. And through the averageroot mean square error analysis, established agent models were all meet the accuracy requirement, where RBF is better than others.Thirdly, based on the sample points and responses, the influences on the system responsescause by the variables changes were researched and analyzed. Then the main effect graphs andthe Pareto graphs were obtained, both of which show that the main factor is the modulus. Theinteraction effect was also analyzed, and the results show any two parameters were interactionno matter to the highest contact stress on the tooth face or to the friction dissipation. Then thecorrelation analysis, local effect, and global effect were also analyzed, and the ranking ofcorrelation local effect, and global effect were obtained too.At last, based on the established agent model, taking minimum the highest contact stresson the tooth face, the friction dissipation, and the gears volume as target, the parameters wereoptimized through multi-island genetic algorithm, non-dominated sorting genetic algorithm-Ⅱ(NSGA-Ⅱ), and multi-objective particle swarm optimization algorithm in multi-subject designoptimization software Isight. During the optimization, the optimization variables were themodulus, tooth number of planet gears, and transmission ratio, while the axial deflection andaxial backoff were restrained. The optimization results show that each of the intelligentalgorithms can find the location of optimized optimization solutions, and the solutions wereclose to each other, and the results of NSGA-Ⅱ are the best.This paper provides a feasible design method for PSD optimization. And the effect lawbetween design variables and system responses was analyzed, which provides evidence forfurther experimental analysis.